Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysis for PsychologistsUsing qgraph in R
Sacha Epskamp
University of AmsterdamDepartment of Psychological Methods
22-05-2014APS 2014
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I These sheets are available onhttp://sachaepskamp.com/presentations
I Including an R script containing codes of the lastsection
This workshop is cosponsored by APS and the Society ofMultivariate Experimental Psychology (SMEP).
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
http://www.psychosystems.org/
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Psychosystems
Staff:I Denny BorsboomI Angelique CramerI Lourens WaldorpI Sacha EpskampI Claudia van BorkuloI Mijke Rhemtulla
Students:I Marie DesernoI Jolanda
KossakowskiI Pia Tio
Collaborators:I Giulio CostantiniI Jonathon LoveI Laura BringmannI Francis TuerlinckxI Laura RuzzanoI Hilde GeurtsI . . .
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
MDInsomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Insomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Insomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Psychology as a complex system
●
●
● ●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●●● ●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●●
●●
●
●
●
●
●
●●
●●
●
●
●
●●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●●
● ● ●
●●
●●●●
●
●●
●
●
●
●●
●●
●
● ●
●
●●
●
●
●
●
●●●
●●
●●
●
●●
●
●
●
●
●
●●
●
●
●
●
●●
●
●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●
●
●●
●
●●
●
●
●●
●●
●
●●
●●
●●
●
●●
●
●●
●●
●
●
●
●●
●●
●
●●●
●
●
●
●
●●● ●
●●
●●●
●
●
●
●
●● ● ●
●
●
●
●
●
●●
●
●
●●
●●
●
●
●
●●
●●
●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
● ●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
● ●
●
●●
●
●
●
●
●
●
●●
● ●●
● ●
●●
● ●
●●
●
●
●●
●
●●
●
● ●●
●
●●
●
●
●●
●
●●
●●
●
●●
● ●●
● ●
●
●
●●●
●
●
●
●
●
●
●
●
●
●●
●
●
●●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
Disorders usually first diagnosed in infancy, childhood or adolescenceDelirium, dementia, and amnesia and other cognitive disordersMental disorders due to a general medical conditionSubstance−related disordersSchizophrenia and other psychotic disordersMood disordersAnxiety disordersSomatoform disordersFactitious disordersDissociative disordersSexual and gender identity disordersEating disordersSleep disordersImpulse control disorders not elsewhere classifiedAdjustment disordersPersonality disordersSymptom is featured equally in multiple chapters
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Edge
Node 1
Node 2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I People
I CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contactI DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI Cities
I SymptomsI An edge represents some connection between two
nodes. Again, this can be anything:
I Friendship / contactI DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contactI DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contactI DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contact
I DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contactI Distance
I Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is a network?
I A network is a set of nodes connected by a set ofedges
I A node represents an entity, this can be anything:I PeopleI CitiesI Symptoms
I An edge represents some connection between twonodes. Again, this can be anything:
I Friendship / contactI DistanceI Comorbidity
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Anne is friends with Laura:
Friendship
Anne
Laura
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Anne is friends with Laura and Roger, but Laura is notfriends with Roger:
Anne
LauraRoger
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Networks can be weightedAnne is better friends with met Roger than Laura:
Anne
LauraRoger
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Weights can be signedAnne is friends with Roger and Laura, but Roger andLaura don’t like each other at all!
Anne
LauraRoger
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Networks can be directedAnne likes Laura, but Laura doesn’t like Anne:
Anne
Laura
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Directed Undirected
Unw
eigh
ted
Wei
ghte
d
●
●
●●
●
●●
●
●●
●
●
●●
●
●●
●
●●
1
2
3
4
5
6
7
8
9
10
●
●
●●
●
●●
●
●●
●
●
●●
●
●●
●
●●
1
2
3
4
5
6
7
8
9
10
●
●
●●
●
●●
●
●●
●
●
●●
●
●●
●
●●
1
2
3
4
5
6
7
8
9
10
●
●
●●
●
●●
●
●●
●
●
●●
●
●●
●
●●
1
2
3
4
5
6
7
8
9
10
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting Networks
A network can be interpreted in different ways:I As a model of interacting components
I Information can spread from node to node via edgesI As a causal modelI As a predictive model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting Networks
A network can be interpreted in different ways:I As a model of interacting components
I Information can spread from node to node via edges
I As a causal modelI As a predictive model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting Networks
A network can be interpreted in different ways:I As a model of interacting components
I Information can spread from node to node via edgesI As a causal model
I As a predictive model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting Networks
A network can be interpreted in different ways:I As a model of interacting components
I Information can spread from node to node via edgesI As a causal modelI As a predictive model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interacting components
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Psychopathology as a virus...
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Networks can model causalityA motivated student will get higher grades:
Motivated
Grade
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
A difficult class causes students to get lower grades:
DifficultClass
Grade
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Motivated
Grade
DifficultClass
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Insomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Causal models
I Methods such as Structural Equation Modeling(SEM) can be used to confirmatory test causalstructures
I Exploratory inference for causal structures is muchharder and mostly not possible without experimentaldesign
I In cross-sectional data only specific structures underthe strong assumption of acyclicness are inferable
I time series could be used to test Granger-causality,but this is again only a confirmatory test
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Causal models
I Methods such as Structural Equation Modeling(SEM) can be used to confirmatory test causalstructures
I Exploratory inference for causal structures is muchharder and mostly not possible without experimentaldesign
I In cross-sectional data only specific structures underthe strong assumption of acyclicness are inferable
I time series could be used to test Granger-causality,but this is again only a confirmatory test
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Causal models
I Methods such as Structural Equation Modeling(SEM) can be used to confirmatory test causalstructures
I Exploratory inference for causal structures is muchharder and mostly not possible without experimentaldesign
I In cross-sectional data only specific structures underthe strong assumption of acyclicness are inferable
I time series could be used to test Granger-causality,but this is again only a confirmatory test
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Causal models
I Methods such as Structural Equation Modeling(SEM) can be used to confirmatory test causalstructures
I Exploratory inference for causal structures is muchharder and mostly not possible without experimentaldesign
I In cross-sectional data only specific structures underthe strong assumption of acyclicness are inferable
I time series could be used to test Granger-causality,but this is again only a confirmatory test
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I A network we can infer is a predictive network
I In predictive networks, we draw an edge from node Ato node B if node A predicts node B.
I Underlying causal networks can be directlytranslated into predictive networks
I Equivalent causal networks lead to the samepredictive network
I If a interaction network generated the data thepredictive network can correctly estimate thestructure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I A network we can infer is a predictive networkI In predictive networks, we draw an edge from node A
to node B if node A predicts node B.
I Underlying causal networks can be directlytranslated into predictive networks
I Equivalent causal networks lead to the samepredictive network
I If a interaction network generated the data thepredictive network can correctly estimate thestructure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I A network we can infer is a predictive networkI In predictive networks, we draw an edge from node A
to node B if node A predicts node B.I Underlying causal networks can be directly
translated into predictive networks
I Equivalent causal networks lead to the samepredictive network
I If a interaction network generated the data thepredictive network can correctly estimate thestructure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I A network we can infer is a predictive networkI In predictive networks, we draw an edge from node A
to node B if node A predicts node B.I Underlying causal networks can be directly
translated into predictive networksI Equivalent causal networks lead to the same
predictive network
I If a interaction network generated the data thepredictive network can correctly estimate thestructure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I A network we can infer is a predictive networkI In predictive networks, we draw an edge from node A
to node B if node A predicts node B.I Underlying causal networks can be directly
translated into predictive networksI Equivalent causal networks lead to the same
predictive networkI If a interaction network generated the data the
predictive network can correctly estimate thestructure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
I Does A predict B?
I Yes!I Does B predict A?
I Yes!I If node A predicts node B, node B predicts node A.
Hence, predictive networks often are undirectednetworks
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
I Does A predict B?I Yes!
I Does B predict A?I Yes!
I If node A predicts node B, node B predicts node A.Hence, predictive networks often are undirectednetworks
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
I Does A predict B?I Yes!
I Does B predict A?
I Yes!I If node A predicts node B, node B predicts node A.
Hence, predictive networks often are undirectednetworks
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
I Does A predict B?I Yes!
I Does B predict A?I Yes!
I If node A predicts node B, node B predicts node A.Hence, predictive networks often are undirectednetworks
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
I Does A predict B?I Yes!
I Does B predict A?I Yes!
I If node A predicts node B, node B predicts node A.Hence, predictive networks often are undirectednetworks
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A B
Predictive model:
A B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C or vise-versa?
I A and C will be correlated, thus knowing A allowsyou to predict C
I But if we also know B, A will not add informationabout C.
I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?I Yes!
I Does B predict C?I Yes!
I Does A predict C or vise-versa?I A and C will be correlated, thus knowing A allows
you to predict CI But if we also know B, A will not add information
about C.I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?I Yes!
I Does B predict C?
I Yes!I Does A predict C or vise-versa?
I A and C will be correlated, thus knowing A allowsyou to predict C
I But if we also know B, A will not add informationabout C.
I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?I Yes!
I Does B predict C?I Yes!
I Does A predict C or vise-versa?I A and C will be correlated, thus knowing A allows
you to predict CI But if we also know B, A will not add information
about C.I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?I Yes!
I Does B predict C?I Yes!
I Does A predict C or vise-versa?
I A and C will be correlated, thus knowing A allowsyou to predict C
I But if we also know B, A will not add informationabout C.
I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
I Does A predict B?I Yes!
I Does B predict C?I Yes!
I Does A predict C or vise-versa?I A and C will be correlated, thus knowing A allows
you to predict CI But if we also know B, A will not add information
about C.I Thus A only predicts C via B
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I In an association network nodes are connected ifthey predict each other regardless of other nodes
I Edges represent correlations
I In a concentration network nodes are connected ifthey predict each other given that we know othernodes
I Edges represent partial correlations or multipleregression coefficients
I If there is no edge between two nodes they areconditionally independent given all other nodes
I Such networks are also called Markov RandomFields.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I In an association network nodes are connected ifthey predict each other regardless of other nodes
I Edges represent correlationsI In a concentration network nodes are connected if
they predict each other given that we know othernodes
I Edges represent partial correlations or multipleregression coefficients
I If there is no edge between two nodes they areconditionally independent given all other nodes
I Such networks are also called Markov RandomFields.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I In an association network nodes are connected ifthey predict each other regardless of other nodes
I Edges represent correlationsI In a concentration network nodes are connected if
they predict each other given that we know othernodes
I Edges represent partial correlations or multipleregression coefficients
I If there is no edge between two nodes they areconditionally independent given all other nodes
I Such networks are also called Markov RandomFields.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I In an association network nodes are connected ifthey predict each other regardless of other nodes
I Edges represent correlationsI In a concentration network nodes are connected if
they predict each other given that we know othernodes
I Edges represent partial correlations or multipleregression coefficients
I If there is no edge between two nodes they areconditionally independent given all other nodes
I Such networks are also called Markov RandomFields.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I In an association network nodes are connected ifthey predict each other regardless of other nodes
I Edges represent correlationsI In a concentration network nodes are connected if
they predict each other given that we know othernodes
I Edges represent partial correlations or multipleregression coefficients
I If there is no edge between two nodes they areconditionally independent given all other nodes
I Such networks are also called Markov RandomFields.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better grades
I A student with good grades is more likely to finishcollege
I IQ is correlated with finishing college and thuspredicts finishing college
I Association
I Given that we know the grades a student got, the IQscore does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better gradesI A student with good grades is more likely to finish
college
I IQ is correlated with finishing college and thuspredicts finishing college
I Association
I Given that we know the grades a student got, the IQscore does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better gradesI A student with good grades is more likely to finish
collegeI IQ is correlated with finishing college and thus
predicts finishing college
I AssociationI Given that we know the grades a student got, the IQ
score does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better gradesI A student with good grades is more likely to finish
collegeI IQ is correlated with finishing college and thus
predicts finishing collegeI Association
I Given that we know the grades a student got, the IQscore does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better gradesI A student with good grades is more likely to finish
collegeI IQ is correlated with finishing college and thus
predicts finishing collegeI Association
I Given that we know the grades a student got, the IQscore does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
IQ
Grades
Finishingcollege
I A smart student will get better gradesI A student with good grades is more likely to finish
collegeI IQ is correlated with finishing college and thus
predicts finishing collegeI Association
I Given that we know the grades a student got, the IQscore does not improve the prediction of finishingcollege
I Concentration
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
A
B
C
Association network:
A
B
C
Concentration network:
A
B
C
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C?
I Yes!If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!
I Does B predict C?I Yes!
I Does A predict C?I Yes!
If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C?
I Yes!If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!
I Does A predict C?I Yes!
If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C?
I Yes!If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C?
I Yes!
If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Where F is a unobservable latent variableI Does A predict B?
I Yes!I Does B predict C?
I Yes!I Does A predict C?
I Yes!If a latent cause underlies the nodes, all nodes predict allother nodes. Since we can not condition on F theprediction is not mediated. The association andconcentration networks then are the same: fullyconnected clusters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Association network:
A
B
C
Concentration network:
A
B
C
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True generating model:
F
A B C
Association network:
A
B
C
Concentration network:
A
B
C
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I Easily constructable
I Make no assumptions about directionalityI Naturally cyclicI Especially edges concentration networks can identify
causal effectsI Correspond to undirected network that could underlie
the data generating model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I Easily constructableI Make no assumptions about directionality
I Naturally cyclicI Especially edges concentration networks can identify
causal effectsI Correspond to undirected network that could underlie
the data generating model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I Easily constructableI Make no assumptions about directionalityI Naturally cyclic
I Especially edges concentration networks can identifycausal effects
I Correspond to undirected network that could underliethe data generating model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I Easily constructableI Make no assumptions about directionalityI Naturally cyclicI Especially edges concentration networks can identify
causal effects
I Correspond to undirected network that could underliethe data generating model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Predictive networks
I Easily constructableI Make no assumptions about directionalityI Naturally cyclicI Especially edges concentration networks can identify
causal effectsI Correspond to undirected network that could underlie
the data generating model
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributed
I For ordinal data threshold models can be used(These can easily be computed in qgraph)
I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu,Lafferty, & Wasserman, 2009)
I For binary data we can use the Ising Model
I To start, an association network for cross-sectionaldata is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrix
I We will assume that all nodes are normallydistributed
I For ordinal data threshold models can be used(These can easily be computed in qgraph)
I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising Model
I To start, an association network for cross-sectionaldata is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributed
I For ordinal data threshold models can be used(These can easily be computed in qgraph)
I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)
I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)I Polychoric correlations
I Polyserial correlationsI nonnormal continuous data could be transformed
using the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising Model
I To start, an association network for cross-sectionaldata is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Estimating Network Topology
I In this block we will look at how to estimateassociation and concentration networks forcross-sectional and time-series data.
I All these methods only require a covariance matrixI We will assume that all nodes are normally
distributedI For ordinal data threshold models can be used
(These can easily be computed in qgraph)I Polychoric correlationsI Polyserial correlations
I nonnormal continuous data could be transformedusing the nonparanormal transformation (Liu et al.,2009)
I For binary data we can use the Ising ModelI To start, an association network for cross-sectional
data is simply a graph of the correlation matrix
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Represent each variable with a node:
1
2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Connect them with a green edge if they are positivelycorrelated:
0.5
1
2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Or a red edge if they are negatively correlated:
−0.5
1
2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The weaker the correlation the vaguer/thinner the edge:
0.2
1
2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Real data example
Included in qgraph is a dataset in which the Dutchtranslation of a commonly used personality test, theNEO-PI-R (Costa & McCrae, 1992; Hoekstra, de Fruyt, &Ormel, 2003), was administered to 500 first yearpsychology students (Dolan, Oort, Stoel, & Wicherts,2009). The NEO-PI-R consists of 240 items designed tomeasure the five central personality factors:
I NeuroticismI ExtroversionI AgreeablenessI Openness to ExperienceI Conscientiousness
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
●
●
●●
●
N1
E2
O3A4
C5
N6
E7
O8A9
C10
N11
E12
O13A14
C15
N16
E17
O18A19
C20
N21
E22
O23A24
C25
N26
E27
O28A29
C30
N31
E32
O33A34
C35
N36
E37
O38A39
C40
N41
E42
O43A44
C45
N46
E47
O48A49
C50
N51
E52
O53A54
C55
N56
E57
O58A59
C60
N61
E62
O63A64
C65
N66
E67
O68A69
C70
N71
E72
O73A74
C75
N76
E77
O78A79
C80
N81
E82
O83A84
C85
N86
E87
O88A89
C90
N91
E92
O93A94
C95
N96
E97
O98A99
C100
N101
E102
O103A104
C105
N106
E107
O108A109
C110
N111
E112
O113A114
C115
N116
E117
O118A119
C120
N121
E122
O123A124
C125
N126
E127
O128A129
C130
N131
E132
O133A134
C135
N136
E137
O138A139
C140
N141
E142
O143A144
C145
N146
E147
O148A149
C150
N151
E152
O153A154
C155
N156
E157
O158A159
C160
N161
E162
O163A164
C165
N166
E167
O168A169
C170
N171
E172
O173A174
C175
N176
E177
O178A179
C180
N181
E182
O183A184
C185
N186
E187
O188A189
C190
N191
E192
O193A194
C195
N196
E197
O198A199
C200
N201
E202
O203A204
C205
N206
E207
O208A209
C210
N211
E212
O213A214
C215
N216
E217
O218A219
C220
N221
E222
O223A224
C225
N226
E227
O228A229
C230
N231
E232
O233A234
C235
N236
E237
O238A239
C24
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
Big 5 correlations
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
● ●●
●
●
●
●
●
●
●
● ●
●
●
●
●●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●●
●●
●
●●
●●
●
●
●
●
●
●
●
● ●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
2627
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
7677
78
79
80
81
82
83
84
85
86
87
88
89
90
9192
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116 117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216217
218
219
220
221222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239240
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
●
●
●
●
●
NeuroticismExtraversionOpennessAgreeablenessConscientiousness
Big 5 correlations
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Toy dataset
Worry (W) Concentration (C) Fatigue (F) Insomnia (I)4 2 3 11 6 2 15 4 4 64 2 8 82 9 1 32 6 2 58 2 4 81 5 4 14 4 5 34 2 5 4
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Insomnia
Fatigue
Concentration
Worry
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Data
## W C F I## 1 4 2 3 1## 2 1 6 2 1## 3 5 4 4 6## 4 4 2 8 8## 5 2 9 1 3## 6 2 6 2 5## 7 8 2 4 8## 8 1 5 4 1## 9 4 4 5 3## 10 4 2 5 4
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association
round(cor(Data),2)
## W C F I## W 1.00 -0.67 0.40 0.70## C -0.67 1.00 -0.73 -0.38## F 0.40 -0.73 1.00 0.52## I 0.70 -0.38 0.52 1.00
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association
W C F IWorry 1.00Concentration -0.67 1.00Fatigue 0.40 -0.73 1.00Insomnia 0.70 -0.38 0.52 1.00
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
AssociationW C F I
Worry 1.00Concentration -0.67 1.00Fatigue 0.40 -0.73 1.00Insomnia 0.70 -0.38 0.52 1.00
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association
library("qgraph")qgraph(cor(Data))
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association networks...I Allow for a powerful visualization of the correlational
structure
I Indicates the presence of latent variablesI But also include many spurious connectionsI And thus not well suited for network analysis
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association networks...I Allow for a powerful visualization of the correlational
structureI Indicates the presence of latent variables
I But also include many spurious connectionsI And thus not well suited for network analysis
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association networks...I Allow for a powerful visualization of the correlational
structureI Indicates the presence of latent variablesI But also include many spurious connections
I And thus not well suited for network analysis
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Association networks...I Allow for a powerful visualization of the correlational
structureI Indicates the presence of latent variablesI But also include many spurious connectionsI And thus not well suited for network analysis
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can construct a concentration network by predictingall variables from all other variables:
W = β21C + β31F + β41I + εW
Where εW ∼ N(0, σW ) and is assumed to be independentof all variables that are not W
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can construct a concentration network by predictingall variables from all other variables:
W = β21C + β31F + β41I + εW
Where εW ∼ N(0, σW ) and is assumed to be independentof all variables that are not W
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can construct a concentration network by predictingall variables from all other variables:
W = β21C + β31F + β41I + εW
Where εW ∼ N(0, σW ) and is assumed to be independentof all variables that are not W
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can construct a concentration network by predictingall variables from all other variables:
W = β21C + β31F + β41I + εW
Where εW ∼ N(0, σW ) and is assumed to be independentof all variables that are not W
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can construct a concentration network by predictingall variables from all other variables:
W = β21C + β31F + β41I + εW
Where εW ∼ N(0, σW ) and is assumed to be independentof all variables that are not W
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
fitW <- lm(W ~ C + F + I, data = Data)fitW
#### Call:## lm(formula = W ~ C + F + I, data = Data)#### Coefficients:## (Intercept) C F## 6.606 -0.723 -0.564## I## 0.518
fitC <- lm(C ~ W + F + I, data = Data)fitF <- lm(F ~ W + C + I, data = Data)fitI <- lm(I ~ W + C + F, data = Data)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C+β31F + β41I + εW
C = β12W+β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
The regression coefficients have a symmetricalrelationship:
βijσ2j = βjiσ
2i
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C+β31F + β41I + εW
C = β12W+β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
The regression coefficients have a symmetricalrelationship:
βijσ2j = βjiσ
2i
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
beta_WC <- coef(fitW)[2]beta_CW <- coef(fitC)[2]sigma_W <- summary(fitW)$sigmasigma_C <- summary(fitC)$sigma
beta_WC * sigma_C^2
## C## -1.163
beta_CW * sigma_W^2
## W## -1.163
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C + β31F + β41I + εW
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C + β31F + β41I + εW
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
These regression parameters can be standardized toobtain the partial correlation coefficient:
ρij =βijσj
σi=βjiσi
σj
Through mathematical magic, these correspond directlyto the negative standardized off-diagonal elements of theinverse of the correlation matrix!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C + β31F + β41I + εW
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
These regression parameters can be standardized toobtain the partial correlation coefficient:
ρij =βijσj
σi=βjiσi
σj
Through mathematical magic, these correspond directlyto the negative standardized off-diagonal elements of theinverse of the correlation matrix!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
W = β21C + β31F + β41I + εW
C = β12W + β32F + β42I + εC
F = β13W + β23C + β43I + εF
I = β14W + β24C + β34F + εI
These regression parameters can be standardized toobtain the partial correlation coefficient:
ρij =βijσj
σi=βjiσi
σj
Through mathematical magic, these correspond directlyto the negative standardized off-diagonal elements of theinverse of the correlation matrix!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
beta_WC * sigma_C / sigma_W
## C## -0.7651
beta_CW * sigma_W / sigma_C
## W## -0.7651
-cov2cor(solve(cov(Data)))
## W C F I## W -1.0000 -0.7651 -0.5889 0.7754## C -0.7651 -1.0000 -0.7993 0.5871## F -0.5889 -0.7993 -1.0000 0.6468## I 0.7754 0.5871 0.6468 -1.0000
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
These partial correlations give us the concentrationgraph:
library("qgraph")qgraph(cor(Data), graph = "concentration")
W
C
F
I
Why does this still not look good?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
These partial correlations give us the concentrationgraph:
library("qgraph")qgraph(cor(Data), graph = "concentration")
W
C
F
I
Why does this still not look good?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
n = 10!
W C F I4 2 3 11 6 2 15 4 4 64 2 8 82 9 1 32 6 2 58 2 4 81 5 4 14 4 5 34 2 5 4
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Coefficients for Worry:
Table:
Dependent variable:W
C −0.723∗∗ (0.248)F −0.564 (0.316)I 0.518∗∗ (0.172)Constant 6.606∗∗ (2.057)
Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Coefficients for Concentration:
Table:
Dependent variable:C
W −0.810∗∗ (0.278)F −0.810∗∗ (0.249)I 0.415 (0.234)Constant 8.451∗∗∗ (0.992)
Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Coefficients for Fatigue:
Table:
Dependent variable:F
W −0.615 (0.345)C −0.789∗∗ (0.242)I 0.451∗ (0.217)Constant 7.460∗∗∗ (1.810)
Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Coefficients for Insomnia:
Table:
Dependent variable:I
W 1.161∗∗ (0.386)C 0.830 (0.467)F 0.927∗ (0.446)Constant −7.071 (4.176)
Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman, Hastie, & Tibshirani,2008) is a very fast algorithm.
I Implemented in glasso package (Friedman, Hastie, &Tibshirani, 2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparison
I Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model space
I p-values and NHST are evilI A well accepted approach for jointly model selection
and parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I We are specifically interested in identifying whichpartial correlations are zero
I These elements correspond to missing edges in thegraph
I We could identify zeroes by significance testing orstepwise model-selection, but...
I Problem of multiple comparisonI Intractable model spaceI p-values and NHST are evil
I A well accepted approach for jointly model selectionand parameter estimation is the least absoluteshrinkage and selection operator (LASSO)
I Limits the sum of absolute regression weights, whichcauses insignificant edges to shrink to zero
I Useful in any regression case. For graphs thegraphical lasso (Friedman et al., 2008) is a very fastalgorithm.
I Implemented in glasso package (Friedman et al.,2011)
I Even usable when you have more variables thenmeasures!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0
## With rho=0, there may be convergence problems if the input matrix s## is not of full rank
0.59−0.59
0.65
−0.770.78
−0.8
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0.1
−0.320.33
0.44
−0.60.64
−0.66
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0.25
−0.140.16
0.31
−0.480.55
−0.57
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0.3
−0.010.03
0.22
−0.390.49
−0.5
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0.4
0.21
−0.370.47
−0.48
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 0.5
0.21
−0.360.45
−0.46
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 1
0.19
−0.31
−0.36
0.39
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 2
0.09
−0.18
−0.2
0.25
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 3
−0.04
−0.05
0.11
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
glasso penalty: 4
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
We can compute an information criterion for each penaltyparameter, and choose the model with the best (lowest)value. In concentration networks the extended BayesianInformation Criterium (EBIC) works well (Foygel & Drton,2010):
EBIC = −2L + |E | log n + 4|E |γ log p
Where L is the log-likelihood, |E | the number of edges inthe graph, n the sample size and p the number ofvariables in the graph. γ is typically set to 0.5 or 0 toobtain regular BIC.In essence this is the likelihood function with a penalty forthe amount of parameters.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Find optimal penalty given extended BIC:
130
135
140
145
0 2 4 6Penalty
EB
IC
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Find optimal penalty given extended BIC:
130
135
140
145
0 2 4 6Penalty
EB
IC
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Find a sparse concentration graph using glasso:
qgraph(cor(Data), graph = "glasso", sampleSize = 10)
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
●●
●●●●●●
●●●● ●●
●●●●
●●●●
●●●●●●
●●
●●●●
●●●●●●
●● ●●●● ●●
Hsi
Hfa
HgaHmo
Efe
EanEde
Ese
Xss
Xsb
Xso
Xli
AfoAge
Afl
Apa
Cor
Cdi
CpeCpr
Oaa Oin
OcrOun
A. Correlation network
●●●●
●●●●
●●●●
●●●●●●
●●●●
●●
●●
●●●●●●
●●●●●●
●●
●● ●●
●●
●●
Hsi
Hfa
Hga
Hmo
Efe
Ean
Ede
EseXss
Xsb
Xso
Xli
Afo
AgeAfl
Apa
Cor
CdiCpe
Cpr
Oaa
Oin
Ocr
Oun
B. Partial Correlation network
●●●●
●●
●●
●●
●● ●●●●
●●
●●●● ●●
●●●●●●
●●
●●
●●●●
●●
●● ●●●● ●●
Hsi
Hfa
Hga
Hmo
Efe
EanEde
Ese
Xss
Xsb
XsoXli
AfoAge
Afl
Apa
Cor
Cdi
Cpe
Cpr
Oaa Oin
Ocr Oun
C. Adaptive lasso network
Paper by Giulio Costantini, Sacha Epskamp, Denny Borsboom, MarcoPerugini, René Mõttus, Lourens J. Waldorp & Angelique O. J. Cramersubmitted
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Partial correlations (using adaptive lasso)
Paper by Jolanda J. Kossakowski, Jacobien M. Kieffer, SachaEpskamp & Denny Borsboom in preparation
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Zooming in: Physical Functioning en Item 2
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
If the data is binary we can use the Ising model to estimate aconcentration network. We have implemented a similar method asdescribed in these slides for continuous data in the IsingFit package(van Borkulo, Epskamp, & with contributions fromAlexander Robitzsch, 2014)
fal slewak
hyp
sad
irr
anx
rmo
mod
qmo
app
wei
con
sel
fut
sui
int
eneple
sexslo
resach
bod
pan
dig
sen
par
Paper by Claudia D. van Borkulo, Denny Borsboom, Sacha EpskampTessa F. Blanken, Lynn Boschloo, Robert A. Schoevers & Lourens J.Waldorp submitted
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Time series
I In time-series data, we do not want to predict everynode given every other node
I It is senseless to predict the past from the future, orpredict the future given other variables in the future
I We don’t care how well a patient being suicidal todaypredicts how depressed he was yesterday
I We don’t care how well a patient being depressedtomorrow will predict how suicidal he will be tomorrow
I Therefore we only set up a predictive model betweentime points, which can be condensed in a directednetwork
I Looking only between consecutive time points iscalled a Lag-1 process. We will not look at morecomplicated models today
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
True causal model between days
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Simulated for 100 days:
−4
−2
0
2
0 25 50 75 100Time
Val
ue
Variable
Worry
Concentration
Fatigue
Insomnia
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
head(Data)
## W C F I## 1 0.0 0.0 0.0 0.0## 2 -0.6 0.2 -0.8 1.6## 3 0.0 -0.8 0.5 1.4## 4 0.8 -0.6 2.1 1.1## 5 -0.1 -2.2 2.4 0.7## 6 0.5 0.5 2.2 0.9
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
We can compute the correlations between consecutivetime-points:
corL1 <- cor(Data[-nrow(Data),],Data[-1,])round(corL1,2)
## W C F I## W 0.50 -0.21 0.22 0.42## C -0.33 0.40 -0.07 -0.21## F 0.14 0.18 0.49 0.18## I 0.10 0.02 0.44 0.58
Which gives us the Lag-1 correlations, and correspond tothe association network
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Lag-1 correlations
qgraph(corL1, layout = "circle")
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
A subject over time...
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concentration
We can again construct a concentration network byrepeated multiple regressions. This time however weregress on the previous time point
Wt = β11Wt−1 + β21Ct−1 + β31Ft−1 + β41It−1 + εW
Ct = β12Wt−1 + β22Ct−1 + β32Ft−1 + β42It−1 + εC
Ft = β13Wt−1 + β23Ct−1 + β33Ft−1 + β43It−1 + εF
It = β14Wt−1 + β24Ct−1 + β34Ft−1 + β44It−1 + εI
Where εi ∼ N(0, σi)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
fitW <- lm(W[-1] ~ W[-100] + C[-100] + F[-100] + I[-100], data = Data)fitW
#### Call:## lm(formula = W[-1] ~ W[-100] + C[-100] + F[-100] + I[-100], data = Data)#### Coefficients:## (Intercept) W[-100] C[-100]## -0.1490 0.4450 -0.2455## F[-100] I[-100]## 0.0629 -0.0573
fitC <- lm(C[-1] ~ W[-100] + C[-100] + F[-100] + I[-100], data = Data)fitF <- lm(F[-1] ~ W[-100] + C[-100] + F[-100] + I[-100], data = Data)fitI <- lm(I[-1] ~ W[-100] + C[-100] + F[-100] + I[-100], data = Data)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
This model is called a Vector-autoregression (VAR) model
y t = By t−1 + ε
In qgraph the VARglm() function can be used toestimate the regression parameters. The transpose of theestimates correspond to the graph structure (for now).Model selection and LASSO can be used, but have notyet been implemented automatically in qgraph
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
VAR network
Beta <- VARglm(Data)$graphqgraph(t(Beta), layout = "circle", labels = names(Data))
W
C
F
I
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
C
E
D
HT
F
P
!"#$%&%! + "' (')$*'%+ "+ (+)$*' + ... + #!
Bringmann, L., Vissers, N., Wichers, M., Geschwind, N., Kuppens, P., Peeters, F., Borsboom, D., & Tuerlinckx, F. (2013). A network approach to psychopathology: New insights into clinical longitudinal data. PLoS ONE.
Wednesday, October 2, 13
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network estimation
Cross-sectional Time-seriesAssociation Correlations Partial correlations∗
Concentration Lag-1 correlations VAR∗
* model selection / LASSO can be used to estimate sparse structure
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I Distance
I How long does it take for a node to influence anothernode?
I Centrality
I Which node is the most important?
I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I DistanceI How long does it take for a node to influence another
node?
I Centrality
I Which node is the most important?
I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I DistanceI How long does it take for a node to influence another
node?I Centrality
I Which node is the most important?I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I DistanceI How long does it take for a node to influence another
node?I Centrality
I Which node is the most important?
I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I DistanceI How long does it take for a node to influence another
node?I Centrality
I Which node is the most important?I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Methodology
Besides providing a interpretable structure of a complexsystem, we can also use networks to compute uniquemeasures such as:
I DistanceI How long does it take for a node to influence another
node?I Centrality
I Which node is the most important?I Connectivity
I How well are nodes connected?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I Each edge can be interpreted as having a length
I Length is inversely related to weight
I Two strongly connected nodes are closer: it is easierfor information to go from one node to the other
I The path between two unconnected nodes isinfinitely long: you cannot walk directly from onenode to the other
I In weighted networks the absolute value of edgeweights is used in computing the length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I Each edge can be interpreted as having a lengthI Length is inversely related to weight
I Two strongly connected nodes are closer: it is easierfor information to go from one node to the other
I The path between two unconnected nodes isinfinitely long: you cannot walk directly from onenode to the other
I In weighted networks the absolute value of edgeweights is used in computing the length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I Each edge can be interpreted as having a lengthI Length is inversely related to weight
I Two strongly connected nodes are closer: it is easierfor information to go from one node to the other
I The path between two unconnected nodes isinfinitely long: you cannot walk directly from onenode to the other
I In weighted networks the absolute value of edgeweights is used in computing the length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I Each edge can be interpreted as having a lengthI Length is inversely related to weight
I Two strongly connected nodes are closer: it is easierfor information to go from one node to the other
I The path between two unconnected nodes isinfinitely long: you cannot walk directly from onenode to the other
I In weighted networks the absolute value of edgeweights is used in computing the length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
I Each edge can be interpreted as having a lengthI Length is inversely related to weight
I Two strongly connected nodes are closer: it is easierfor information to go from one node to the other
I The path between two unconnected nodes isinfinitely long: you cannot walk directly from onenode to the other
I In weighted networks the absolute value of edgeweights is used in computing the length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The distance between Storrs and San Francisco is2638.29 miles (4245.80 km)
2638.29 miles StorrsSan Francisco
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Amsterdam is much farther from San Francisco:
5449.02 mile3523.18 mile
2638.29 mile StorrsSan Francisco
Amsterdam
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
We can place nodes however we want:
5449.02 mile
3523.18 mile 2638.29 mile
Storrs
San FranciscoAmsterdam
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Shortest path length
I The distance between two nodes is defined by thelength of the shortest possible path between twonodes
I In unweighted graphs this is simply the amount ofedges on a path
I In weighted graphs the length of a path is computedas the sum of all edge lengths on a path
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Shortest path length
I The distance between two nodes is defined by thelength of the shortest possible path between twonodes
I In unweighted graphs this is simply the amount ofedges on a path
I In weighted graphs the length of a path is computedas the sum of all edge lengths on a path
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Shortest path length
I The distance between two nodes is defined by thelength of the shortest possible path between twonodes
I In unweighted graphs this is simply the amount ofedges on a path
I In weighted graphs the length of a path is computedas the sum of all edge lengths on a path
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Which node is the most influential?
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodes
I ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strength
I In concentration network the node with the higheststrength directly predicts the most other nodes
I ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodes
I ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodes
I ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodes
I ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodesI ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I Closeness
I In concentration networks the node that (indirectly)has the best predictive value on all other nodes
I ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodes
I ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodesI ...it connects other nodes
I BetweennessI ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodesI ...it connects other nodes
I Betweenness
I ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodesI ...it connects other nodes
I BetweennessI ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Centrality
A node is central/important/influential if...I ...it has many connections
I Degree / strengthI In concentration network the node with the highest
strength directly predicts the most other nodesI ...it is close to all other nodes
I ClosenessI In concentration networks the node that (indirectly)
has the best predictive value on all other nodesI ...it connects other nodes
I BetweennessI ...it is connected to other important nodes
I Eigenvector centrality
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Degree
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
insomnia / difficu...psychomotor agitat...
psychomotor retard...
depressed mood
tachycardia / acce...
often easily distr...
is often touchy or...
anxietynausua
sweating / perspir...Weight loss
difficulty concent...
transient visual, ...
fatigue / fatigue ...
increased appetite
Clinically signifi...
delusions
coma
extreme negativism
wei
ght g
ain
Hypersomnia
feelings of worthl...
Loss of appetite
excessive motor ac...peculiarities of v...
mar
kedl
y di
min
ishe
...
echolalia
echopraxia
recurrent thoughts...
disorganized speech
flight of ideas or...
disorganized behav...
inflated self−este...decreased need for...
more talkative tha...
increase in goal−d...
excessive involvem...
A d
istin
ct p
erio
d ...
motoric immobilitypupillary dilation
mutism
derealization
depersonalization
palpitations
vomiting
fear of one or mor...nystagmus / vertic...
pounding heart
sensations of shor...
abdominal distressThe situations are...
chills or hot flus...
There is evidence ...
fear of losing con...
stupor
elevated blood pre... trembling or shaking
ches
t pai
n or
dis
c...
fear of dying
paresthesias
persistent concern...
worry about the im...
a significant chan...
incoordination
memory impairment ...
The person recogni...
Marked and persist...
rambling flow of t...
slur
red
spee
ch
blurring of vision
tremors
Anxiety about bein...
A distinct period ...
often lies to obta...
often bullies, thr...
often initiates ph...
has used a weapon ... has been physicall...
has been physicall...
has stolen while c...
has forced someone...
has deliberately e...
has deliberately d...
has broken into so...
has stolen items o...
often stays out at...
has run away from ...is often truant fr...
The development of...
often fails to giv...
often has difficul...
often does not see...
ofte
n do
es n
ot fo
l...
often has diff icu...
often avoids, disl...
often loses t hing...is often forgetful...
often fidgets with...
often leaves seat ...
often runs about o...
often has difficul...
is often "on the g...
often talks excess...
often blurts out a...
often has difficul...
often interrupts o...
Excessive anxiety ...
Disturbance of con...
unsteady gait
bradycardia
chills
confusion, seizure...
flat or inappropri...
muscular weakness,...
lowered blood pres...
generalized muscle...
the person experie...
dizziness
lethargy
depressed reflexes
euphoria
Presence of two or...
lacks close friend...
cardiac arrhythmia
piloerection
nervousness
excitement
flushed face
diuresis
gastroint estinal ...
mus
cle
twitc
hing
periods of inexhau...
muscle aches
lacr
imat
ion
or r
hi...
diarrhea
yawning
fever
seiz
ures
Focal neurological...
aphasia
apraxia
agnosia disturbance in exe...
numbness or dimini...
ataxia
dysa
rthr
ia
mus
cle
rigid
ity
hyperacusis
Confusion about pe...
The
per
son
finds
i...
mind going blank
mus
cle
tens
ion
odd beliefs or mag...unusual perceptual...
decreased heart rate
vivid, unpleasant ...
suspiciousness or ...
The traumatic even...
transient, stress ...
Marked avoidance o...
ideas of reference
A change in cognit...
development of a p...
Recent ingestion o...
autonomic hyperact...
increased hand tre...
grand mal seizures
feelings of hopele...
dissociative amnesia
behavior or appear...
frantic efforts to...
a pa
ttern
of u
nsta
...
impulsivity in at ...
recurrent suicidal...
affective instabil...
chronic fee lings ...
failure to conform...deceitfulness
Per
sist
ent a
void
an...
often loses temperoften argues with ...
ofte
n de
liber
atel
y...
often blames other...
is often angry and...
is often spiteful ...
hypervigilance
exaggerated startl...
impulsivity or fai...
reckless disregard...
consistent irrespo...
lack of remorse
often actively def... Dysphoric mood
drowsiness
neither desires no...
has little, if any...
takes pleasure in ...
Pupillary constric...
catalepsy
almost always choo...
appears indifferen...
conjunctival injec...
dry
mou
th
Relative unrespons...
No detailed dream ...
Rec
urre
nt e
piso
des.
..
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Closeness
●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●●
●● ●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●
●●●●
●●●●
●●●●
●● ●●
●●
●●●●
●●●●
●●
●●●●
●●
●●
●●●●
●●
●●
●●●●
●●
●●●●●●●●
●●●●
●●
●●●●●●
●●
●●
●●
●●●●●●
●●
●●●●
●●
●●
●●●●
●●●●
●●
●● ●●
●●●●●●
●● ●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●
●●●●
●●●●
●●●●●●●●●●
●●
●●
●●
●●
●●
●●
●●●●
●●●●
●●
●●●●
●●●●
●●●●
●●●●
●●
●●
●●●●●●
●●●●
●●
●●
●● ●●●●
●●●●
●●
●●
●●
●●
●●
●●●●●●●●
●●
●●
●●
●●
●●
●●
●●
●●
●●●● ●●
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
2526
27
28
2930
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
656667
68
69
70
71
72
73
74
75
76
77
78
79
8081
82
83
8485
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103104
105
106
107
108
109
110111
112
113
114
115
116117
118
119
120
121
122
123
124
125
126
127
128
129130
131
132133
134135
136137
138139
140
141
142143
144
145
146
147
148
149
150
151152
153
154155
156
157
158
159
160
161
162
163
164
165
166
167
168
169170
171
172
173174
175
176
177178
179
180
181
182
183
184 185
186
187
188
189
190
191
192
193
194195
196
197
198
199
200
201
202
203
204
205
206
207208
insomnia / difficu...psychomotor agitat...
increased appetitedepressed mood
psychomotor retard...
nausua
transient visual, ...
difficulty concent...
fatigue / fatigue ...
Weight loss
weight gain
sweating / perspir...is often touchy or...
anxiety
tach
ycar
dia
/ acc
e...
vomiting
often easily distr...
pupillary dilation
delusions
disorganized speech
diso
rgan
ized
beh
av...
extreme negativism
Hypersomnia
feel
ings
of w
orth
l...
echolalia
echopraxia
mutism
com
a
palpitations
incoordination
mus
cle
ache
s
diarrheayawning
fever chills
mus
cula
r w
eakn
ess,
...
tremors
excitement
flush
ed fa
ce
stupor
leth
argy
ataxia
unexpected travel ...
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Small world
The famous paper of Watts and Strogatz (1998)—alreadycited 22691 times—describes the “small world” principlethat frequently occurs in natural graphs.
I “Six degrees of separation”I High clustering and low average path length
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
mouseover for friend details
Hbased on data from 190 of 203 friendsL
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Small world
A graph exhibits a small world if it has both highclustering and a short average path length. Thisinformation can be summarized in a small-world index,which should be higher than 3.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Network Analysis with qgraph
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is R?
I R is a statistical programming languageI Statistical analysisI Data visualizationI Data miningI General programming
I It is open-sourceI Free as in “free beer” and “free speech”I Large active community around RI Many contributed packages
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
What is R?
R can be downloaded fromhttp://cran.r-project.org/. A good environmentfor R is RStudio which can be downloaded fromhttps://www.rstudio.com/. Some good links tostart learning R are:
I http://cran.r-project.org/doc/contrib/Torfs+Brauer-Short-R-Intro.pdf
I https://www.codeschool.com/courses/try-r
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
R uses functions, such as rnorm(), which take somearguments between the parantheses. All these functionsare documented using ?. For example, rnorm generatesrandom numbers:
# rnorm samples random normal numbers:rnorm(3)
## [1] 1.59291 0.04501 -0.71513
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Multiple arguments can be used. They can also benamed. See the documentation for a list of arguments!?rnorm
# Using arguments we can change its behavior. e.g.,rnorm(3, mean = 100, sd = 15)
## [1] 113.0 116.1 128.4
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The <- symbol can be used to store results and reusethem later.
# We can use '<-' to store the result:result <- rnorm(3)result
## [1] -0.6030 -0.3909 -0.4162
# Reuse:mean(result)
## [1] -0.47
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Packages
I Packages are extensions contributed to R containingextra functions
I They can be installed using install.packages()
I Afterwards they can be loaded using library()
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Packages
# Install packageinstall.packages("qgraph", dep = TRUE)
# Load packagelibrary("qgraph")
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Load data into R
I There are many ways to load data into RI The most common way is to read a plain text file
(which can be exported from excel for instance)using read.table or read.csv (see their helpfiles for how to do this)
I Because psychologists often use SPSS, it is usefulto directly import data from SPSS
I This can be done using read.spss() in theforeign package.
I file.choose() can be used to select a file (itmight be opened in the background of RStudio)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Load SPSS data into R
# Load package "foreign":library("foreign")
# Select a SPSS file:file <- file.choose()
# Read data:Data <- read.spss(file,to.data.frame=TRUE)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Rather than using a file as data, we will use a datasetalready included in R via the psysch package (Revelle,2010):
library("psych")data(bfi)bfi <- bfi[,1:25]
str(bfi)
## 'data.frame': 2800 obs. of 25 variables:## $ A1: int 2 2 5 4 2 6 2 4 4 2 ...## $ A2: int 4 4 4 4 3 6 5 3 3 5 ...## $ A3: int 3 5 5 6 3 5 5 1 6 6 ...## $ A4: int 4 2 4 5 4 6 3 5 3 6 ...## $ A5: int 4 5 4 5 5 5 5 1 3 5 ...## $ C1: int 2 5 4 4 4 6 5 3 6 6 ...## $ C2: int 3 4 5 4 4 6 4 2 6 5 ...## $ C3: int 3 4 4 3 5 6 4 4 3 6 ...## $ C4: int 4 3 2 5 3 1 2 2 4 2 ...## $ C5: int 4 4 5 5 2 3 3 4 5 1 ...## $ E1: int 3 1 2 5 2 2 4 3 5 2 ...## $ E2: int 3 1 4 3 2 1 3 6 3 2 ...## $ E3: int 3 6 4 4 5 6 4 4 NA 4 ...## $ E4: int 4 4 4 4 4 5 5 2 4 5 ...## $ E5: int 4 3 5 4 5 6 5 1 3 5 ...## $ N1: int 3 3 4 2 2 3 1 6 5 5 ...## $ N2: int 4 3 5 5 3 5 2 3 5 5 ...## $ N3: int 2 3 4 2 4 2 2 2 2 5 ...## $ N4: int 2 5 2 4 4 2 1 6 3 2 ...## $ N5: int 3 5 3 1 3 3 1 4 3 4 ...## $ O1: int 3 4 4 3 3 4 5 3 6 5 ...## $ O2: int 6 2 2 3 3 3 2 2 6 1 ...## $ O3: int 3 4 5 4 4 5 5 4 6 5 ...## $ O4: int 4 3 5 3 3 6 6 5 6 5 ...## $ O5: int 3 3 2 5 3 1 1 3 1 2 ...
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
This is a dataset measuring the big 5 personality traitsusing an ordinal scale. The qgraph functioncor_auto() can be used to compute an appropriatecorrelation matrix. In this case, polychoric correlations:
cor_bfi <- cor_auto(bfi)
## Variables detected as ordinal: A1; A2; A3; A4;A5; C1; C2; C3; C4; C5; E1; E2; E3; E4; E5; N1; N2;N3; N4; N5; O1; O2; O3; O4; O5
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
We can send this correlation matrix to qgraph() to plotthe association network. To use a force-embeddedalgorithm we set the layout argument to "spring"
graph_cor <- qgraph(cor_bfi, layout = "spring")
A1
A2
A3
A4
A5
C1
C2
C3
C4
C5
E1
E2
E3
E4
E5
N1N2
N3
N4
N5
O1O2
O3
O4
O5
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The partial correlation graph can be plotted by setting thegraph argument to "concentration"
graph_pcor <- qgraph(cor_bfi, graph = "concentration", layout = "spring")
A1
A2A3A4
A5
C1
C2
C3
C4
C5
E1E2
E3
E4
E5
N1
N2
N3
N4
N5
O1
O2
O3O4
O5
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
To use the graphical lasso to estimate a sparse structureset graph to "glasso" and assign the sample size:
graph_glas <- qgraph(cor_bfi, graph = "glasso", sampleSize = nrow(bfi), layout = "spring")
A1
A2 A3A4
A5
C1C2
C3
C4C5
E1
E2
E3
E4
E5
N1
N2
N3
N4 N5 O1
O2
O3O4
O5
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
With legend:
Names <- scan("http://sachaepskamp.com/files/BFIitems.txt",what = "character", sep = "\n")
Groups <- rep(c('A','C','E','N','O'),each=5)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
With legend:
qgraph(cor_bfi, graph = "glasso", sampleSize = nrow(bfi),layout = "spring", nodeNames = Names,legend.cex =0.6, groups = Groups)
A1
A2 A3A4
A5
C1C2
C3
C4C5
E1
E2E3
E4
E5
N1
N2
N3
N4 N5 O1
O2
O3O4
O5
A1: Am indifferent to the feelings of others.A2: Inquire about others' well−being.A3: Know how to comfort others.A4: Love children.A5: Make people feel at ease.C1: Am exacting in my work.C2: Continue until everything is perfect.C3: Do things according to a plan.C4: Do things in a half−way manner.C5: Waste my time.E1: Don't talk a lot.E2: Find it difficult to approach others.E3: Know how to captivate people.E4: Make friends easily.E5: Take charge.N1: Get angry easily.N2: Get irritated easily.N3: Have frequent mood swings.N4: Often feel blue.N5: Panic easily.O1: Am full of ideas.O2: Avoid difficult reading material.O3: Carry the conversation to a higher level.O4: Spend time reflecting on things.O5: Will not probe deeply into a subject.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
A graph can be saved using filetype and filename.PDF files produce the best results:
qgraph(cor_bfi, graph = "glasso",sampleSize = nrow(bfi), layout = "spring",filetype = "pdf", filename = "glasso")
## Output stored in/home/sacha/[email protected]/Documents/Work/Talks/APS/Sheets/glasso.pdf
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The centrality_auto() function automaticallycomputes centrality measures and shortest path lengths:
cent <- centrality_auto(graph_glas)names(cent)
## [1] "node.centrality"## [2] "edge.betweenness.centrality"## [3] "ShortestPathLengths"
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
centralityPlot(graph_glas, labels=Names)
Betweenness Closeness Strength
Am exacting in my work.
Am full of ideas.
Am indifferent to the feelings of others.
Avoid difficult reading material.
Carry the conversation to a higher level.
Continue until everything is perfect.
Don't talk a lot.
Do things according to a plan.
Do things in a half−way manner.
Find it difficult to approach others.
Get angry easily.
Get irritated easily.
Have frequent mood swings.
Inquire about others' well−being.
Know how to captivate people.
Know how to comfort others.
Love children.
Make friends easily.
Make people feel at ease.
Often feel blue.
Panic easily.
Spend time reflecting on things.
Take charge.
Waste my time.
Will not probe deeply into a subject.
0.00 0.25 0.50 0.75 1.00 0.8 0.9 1.00.6 0.7 0.8 0.9 1.0
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
centralityPlot(list(cor = graph_cor, pcor = graph_pcor, glasso = graph_glas), labels=Names)
Betweenness Closeness Strength
Am exacting in my work.
Am full of ideas.
Am indifferent to the feelings of others.
Avoid difficult reading material.
Carry the conversation to a higher level.
Continue until everything is perfect.
Don't talk a lot.
Do things according to a plan.
Do things in a half−way manner.
Find it difficult to approach others.
Get angry easily.
Get irritated easily.
Have frequent mood swings.
Inquire about others' well−being.
Know how to captivate people.
Know how to comfort others.
Love children.
Make friends easily.
Make people feel at ease.
Often feel blue.
Panic easily.
Spend time reflecting on things.
Take charge.
Waste my time.
Will not probe deeply into a subject.
0.000.250.500.751.00 0.7 0.8 0.9 1.0 0.6 0.8 1.0
type
cor
glasso
pcor
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
The smallworldness() function computes thesmallworldness index:
smallworldness(graph_glas)
## smallworldness## 1.0105## trans_target## 0.6358## averagelength_target## 1.3833## trans_rnd_M## 0.6292## trans_rnd_lo## 0.6131## trans_rnd_up## 0.6438## averagelength_rnd_M## 1.3833## averagelength_rnd_lo## 1.3833## averagelength_rnd_up## 1.3833
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Concluding comments
More information is on our website:http://psychosystems.org/ and the developmentalversion of qgraph is on GitHubhttps://github.com/SachaEpskamp/qgraph
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Thank you for your attention!
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
References I
Borsboom, D., Cramer, A. O., Schmittmann, V. D.,Epskamp, S., & Waldorp, L. J. (2011). The smallworld of psychopathology. PloS one, 6(11),e27407.
Costa, P. T., & McCrae, R. R. (1992). Revised NEOPersonality Inventory and NEO Five-FactorInventory. Odessa, FL: PsychologicalAssessment Services.
Cramer, A. O., Waldorp, L. J., van der Maas, H. L.,Borsboom, D., et al. (2010). Comorbidity: Anetwork perspective. Behavioral and BrainSciences, 33(2-3), 137–150.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
References II
Dolan, C., Oort, F., Stoel, R., & Wicherts, J. (2009).Testing Measurement Invariance in the TargetRotated Multigroup Exploratory Factor Model.Structural Equation Modeling: A MultidisciplinaryJournal, 16(2), 20.
Doosje, B., Loseman, A., & Bos, K. (2013). Determinantsof radicalization of islamic youth in thenetherlands: Personal uncertainty, perceivedinjustice, and perceived group threat. Journal ofSocial Issues, 69(3), 586–604.
Foygel, R., & Drton, M. (2010). Extended bayesianinformation criteria for gaussian graphical models.In Nips (pp. 604–612).
Friedman, J., Hastie, T., & Tibshirani, R. (2008). Sparseinverse covariance estimation with the graphicallasso. Biostatistics, 9(3), 432–441.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
References III
Friedman, J., Hastie, T., & Tibshirani, R. (2011). glasso:Graphical lasso- estimation of gaussian graphicalmodels [Computer software manual]. Retrievedfrom http://CRAN.R-project.org/package=glasso (R package version 1.7)
Hoekstra, H., de Fruyt, F., & Ormel, J. (2003).Neo-persoonlijkheidsvragenlijsten: NEO-PI-R,NEO-FFI. Lisse: Swets Test Services.
Liu, H., Lafferty, J., & Wasserman, L. (2009). Thenonparanormal: Semiparametric estimation ofhigh dimensional undirected graphs. The Journalof Machine Learning Research, 10, 2295–2328.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
References IVRevelle, W. (2010). psych: Procedures for psychological,
psychometric, and personality research[Computer software manual]. Evanston, Illinois.Retrieved fromhttp://personality-project.org/r/psych.manual.pdf (R package version 1.0-93)
van Borkulo, C., Epskamp, S., & with contributions fromAlexander Robitzsch. (2014). Isingfit: Fitting isingmodels using the elasso method [Computersoftware manual]. Retrieved from http://CRAN.R-project.org/package=IsingFit(R package version 0.2.0)
Watts, D., & Strogatz, S. (1998). Collective dynamics ofsmall-world networks. Nature, 393(6684),440–442.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Doosje, Loseman, and Bos (2013)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Doosje et al. (2013)
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Correlation network
1
2
3
4
5
6
7
8
9
10
11
12
1314
1: In−group Identification2: Individual Deprivation3: Collective Deprivation4: Intergroup Anxiety5: Symbolic Threat6: Realistic Threat7: Personal Emotional Uncertainty8: Perceived Injustice9: Perceived Illegitimacy authorities10: Perceived In−group superiority11: Distance to Other People12: Societal Disconnected13: Attitude towards Muslim Violence14: Own Violent Intentions
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Partial correlation network
1
2
3
4
5
67
8
9
10
11
12
1314
1: In−group Identification2: Individual Deprivation3: Collective Deprivation4: Intergroup Anxiety5: Symbolic Threat6: Realistic Threat7: Personal Emotional Uncertainty8: Perceived Injustice9: Perceived Illegitimacy authorities10: Perceived In−group superiority11: Distance to Other People12: Societal Disconnected13: Attitude towards Muslim Violence14: Own Violent Intentions
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Partial correlation networkAfter glasso:
1
2
3
4
5
6
7
8
9
10
1112
13
14
1: In−group Identification2: Individual Deprivation3: Collective Deprivation4: Intergroup Anxiety5: Symbolic Threat6: Realistic Threat7: Personal Emotional Uncertainty8: Perceived Injustice9: Perceived Illegitimacy authorities10: Perceived In−group superiority11: Distance to Other People12: Societal Disconnected13: Attitude towards Muslim Violence14: Own Violent Intentions
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Partial correlation networkAfter glasso (Friedman et al., 2011):
1
2
3
4
5
6
7
8
9
10
1112
13
14
1: In−group Identification2: Individual Deprivation3: Collective Deprivation4: Intergroup Anxiety5: Symbolic Threat6: Realistic Threat7: Personal Emotional Uncertainty8: Perceived Injustice9: Perceived Illegitimacy authorities10: Perceived In−group superiority11: Distance to Other People12: Societal Disconnected13: Attitude towards Muslim Violence14: Own Violent Intentions
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Shortest path length
Muslim Violence Violent IntentionsIn-group Identification 16.12 19.89Individual Deprivation 20.10 23.87Collective Deprivation 17.05 20.82
Intergroup Anxiety Inf InfSymbolic Threat 11.44 15.21Realistic Threat 14.81 18.58
Personal Emotional Uncertainty 10.73 14.50Perceived Injustice 24.83 28.60
Perceived Illegitimacy authorities Inf InfPerceived In-group superiority 3.16 6.93
Distance to Other People 5.70 8.81Societal Disconnected 9.15 12.92
Attitude towards Muslim Violence 0.00 3.77Own Violent Intentions 3.77 0.00
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Shortest path length
Muslim Violence Violent IntentionsIn-group Identification 16.12 19.89Individual Deprivation 20.10 23.87Collective Deprivation 17.05 20.82
Intergroup Anxiety Inf InfSymbolic Threat 11.44 15.21Realistic Threat 14.81 18.58
Personal Emotional Uncertainty 10.73 14.50Perceived Injustice 24.83 28.60
Perceived Illegitimacy authorities Inf InfPerceived In-group superiority 3.16 6.93
Distance to Other People 5.70 8.81Societal Disconnected 9.15 12.92
Attitude towards Muslim Violence 0.00 3.77Own Violent Intentions 3.77 0.00
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Extra: Interpreting qgraph
I Under the default coloring scheme, positive edgeweights (here correlations) are shown as greenedges and negative edge weights as red.
I An edge weight of 0 is omitted. The wider and morecolorful an edge the stronger the edge weight.
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
To interpret qgraph networks, three values need to beknown:
Minimum Edges with absolute weights under thisvalue are omitted
Cut If specified, splits scaling of width and colorMaximum If set, edge width and color scale such that
an edge with this value would be the widestand most colorful
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.1 0.1−0.2 0.2−0.3 0.3−0.4 0.4−0.5 0.5−0.6 0.6−0.7 0.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Default settings
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.1 0.1−0.2 0.2−0.3 0.3−0.4 0.4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Default settings
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.1 0.1−0.2 0.2−0.3 0.3−0.4 0.4−0.5 0.5−0.6 0.6−0.7 0.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Maximum 1
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.1 0.1−0.2 0.2−0.3 0.3−0.4 0.4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Maximum 1
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.2 0.2−0.3 0.3−0.4 0.4−0.5 0.5−0.6 0.6−0.7 0.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Minimum 0.1
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.1 0.1−0.2 0.2−0.3 0.3−0.4 0.4−0.5 0.5−0.6 0.6−0.7 0.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Cut 0.4
Network Analysisfor Psychologists
Sacha Epskamp
IntroductionPsychology as a ComplexSystem
Building Blocks of aNetwork
Types of Networks
Interaction Networks
Causal Networks
Predictive Networks
Estimating networktopologyCross-sectional networks
Association networks
Concentration Networks
Time series
NetworkMethodologyDistance and ShortestPaths
Centrality
Connectivity and Clustering
Network Analysiswith qgraphA very tiny introduction to R
Estimating networks inqgraph
Network inference
References
Extra
Interpreting qgraph
−0.2 0.2−0.3 0.3−0.4 0.4−0.5 0.5−0.6 0.6−0.7 0.7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Minimum 0.1Cut 0.4
Maximum 1