Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
1
I t ti O t l M d l d
Protégé Conference Amsterdam – June 2009
Integrating Ontology Models and Conceptual Models
using a Meta Modeling Approach
Hans-Georg Fill, Patrik BurzynskiDepartment of Knowledge and Business Engineering,
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
University of Vienna
Agenda
• Motivation• Conceptual Modeling• Conceptual Modeling• Meta Modeling• Three approaches for an integration• Implementation and Application Scenarios• Outlook
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
2
Motivation
• Several business cases for the use of web-based ontologies:– Enterprise Content Management– Enterprise Information Integration– Enterprise Service Bus– …
• Some common goals:– Let machines manage complexity
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
– Using explicit semantics and reasoning
– Based on shared, web-based,explicit conceptualizations
• How to analyze the contribution to business value?Picture source: http://www.atibatechnology.com/ECM_DM.html
Conceptual Modeling• Conceptual Modeling:
– Supporting humans to analyze organizational and technical relationshipsp
– Based on formal syntax and notation– Optional definition of formal semantics– Goal: Improvement of Human
understanding
• Examples:– Analysis of business models, strategic goals, performance
i i f i
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
measurements, opportunities for action– Management of business processes including representation,
analysis, simulation– …
• How to bring together ontologies and conceptual models?
3
BusinessView
Business Processes
ManagementApproaches
eg. BSC
RegulatoryFrameworks
Business Model
IT-GovernanceFrameworks
Production Model
SemiformalProzeßstart-46430
Aktivität-46433
Name
an
Name
an
Percentage ofemployees
with a flexible Strategic
J
Operatives
J
Kennzahl-42841
Kennzahl-42841
J
Business & IT
SemiformalView
Aktivität-46436
Aktivität-46436-46439
Entscheidung-46442
Aktivität-46451
Aktivität-46454
Aktivität-46457
Name
an
Name
anan
Name
an
Name
an
Name
an
an
Name
an
with a flexibletime schedule
x
/
1100,000000
1 2α
Number ofemployees
with a flexibletime schedule
db
αNumber ofemployees
with atraditional
time schedule
db
StrategicGoal
OperativesZiel-42838 4284142841
StrategischesZiel-42844
J
FormalizedViewB-XML
(e.g. BPEL) S-XML C-XML ……
Technology
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
InfrastructureView
TechnologyView
Services Orchestrated Services Applications… …
Spectrum of „Ontologies“
strongsemantics
First Order LogicModal Logic
Humans usually do not express themselves in a logic-based language…
Thesaurus
Conceptual Model
Logical Theory
ER
Extended ERXTM
RDF/S
UML
DAML+OIL,OWL
Description Logic
Has Narrower Meaning Than
Is Subclass of
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
weak semantics
TaxonomyRelational Model
Schema
Is Sub-Classification of
(Obrst, 2003)
“Ontologies are usually expressed in a logic-based language…”
4
DKE Meta Modeling Framework
modelingmethod
defines grammar
defines meaningModellingLanguagemodelinglanguage
defines visualizationdefines way of language application
deliversmodelingprocedure
modelingtechnique
method
mechanisms& algorithms
used for
used in
genericmechanisms& algorithms
specific
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
arrangesaccording to
Semantics
SemanticSchema
Syntax
SemanticMapping
connectsconsiders
Notation semantics
semanticdomain
syntaxnotation visualizes
semanticmapping
describesmeaning of
steps results
hybridmechanisms& algorithms
mechanisms& algorithms
(Karagiannis et al., 2002)
Meta Modeling Approach
0..* 1..1Ontological VisualizationPattern
External Variable0..* 1..1Ontological Visualization
PatternExternal Variable
Ontological VisualizationPattern
External Variable
Visualization (Notation)Meta Model (Syntax)
BusinessProcess
Flow Element ControlC t
IsFromActivityElement
WorkingEnvironment
WE Element WE Relation
IsFromWEElement
Business Graph
Relations
Personibl f
RequiredR
1..* 1..1
1..1
1..* 1..*
1..1
1..1
1..1
1..*
1..1
1..*
1..1
1..* 1..*
1..* 1..1
1..1 1..1
1..1
1..* 1..*
1..1
0..*0..*
1..*
1..* 0..*
1..*
0..*
1..*consists of contains
linked to
refers to
Visual Object
Graphical Primitive Internal Variable
Transformation Function
refers toconsists of
1..1
1..* 1..*
1..1
0..*0..*
1..*
1..* 0..*
1..*
0..*
1..*consists of contains
linked to
refers to
Visual Object
Graphical Primitive Internal Variable
Transformation FunctionVisual Object
Graphical Primitive Internal Variable
Transformation Function
refers toconsists of
Connector
Activity SubProcessPerson Group Resource
responsible for Resource
responsibleFor
responsibilityBy
requiredFor provisionedBy
IsToActivityElement IsToWEElement1..* 1..11..* 1..1
1..* 1..1
1..11..*
1..*
1..*
1..*
1..*
Control ElementControl Element
StartStart DecisionDecisionEndEnd
Identification of Common Model Characteristics
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
5
Example for the Realization of a Process Meta Model
S
Class Relationclass
C t t
Notify Customer
Customer reached
ActivityDecision
Subsequent
Meta ModelModel
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Contact Customer
Write E-MailCustomer not reached
Semantic Issues
• Meta Models for Conceptual Modeling:– Assumption of implicit human-interpreted semanticsAssumption of implicit, human interpreted semantics– Only „execution“ of models requires strict formalization incl.
formal semantics– Formalization of semantics only for particular purposes, e.g.
simulation– No consideration of inherent semantics of the model content,
e.g. an activity in a process is named „print report“ but no information about print‘ or report‘ is made explicit
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
information about ‚print or ‚report is made explicit– IT-based communication of semantic information requires
common semantic base, e.g. an ontology
How to make semantics explicit?
6
Integrating Explicit Semantics
Three Approaches:Three Approaches:• Integrating Ontologies on the Meta Model Level• Defining References between Meta Model Elements
and externally kept Ontologies, resp. between Model Elements and Ontologies
• Using a Combination of Integration and External Li k
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Linkage
Example for the Realization of an Ontology Model
O C
Class Attribute
OWL ClassProperty
Domain
Meta ModelModel
C C C C
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
foaf:Document dcmitype:Image skos:ConceptScheme skos:Collection
PaltLabel
PaltSymbol
PchangeNote
Pdc:title
7
Integrating Ontologies on the Meta Model Level
Linkages on the Meta Model LevelMeta Model Ontology Meta Model
Contact Customer
Notify Customer
Write E-Mail
Customer reached
Customer t h d
ActivityDecision
Subsequent
Class Relationclass
Meta Model
Model
OWL ClassProperty
Domain
Class Attribute
Meta Model
Model
Cfoaf:Document
Cdcmitype:Image
Cskos:ConceptScheme
Cskos:Collection
P P P P
Meta Model Ontology Meta Model
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
not reachedaltLabel altSymbol changeNote dc:title
Advantages:• Coherent conceptual approach for both ontologies and domain aspects• Inherent graphical representation of ontology• General mechanisms and algorithms for meta models directly applicable to ontologies
Disadvantages:• Graphical modeling of large ontologies is difficult• Visualization of large and complex ontologies difficult to handle for users• Specific ontology aspects might need to be re-implemented / re-configured
References between (Meta) Models and External Ontologies
Meta Model Ontology Meta Model
Contact Customer
Notify Customer
Write E-Mail
Customer reached
Customer not reached
ActivityDecision
Subsequent
Class Relationclass
Meta Model
Model
Extensions Repository
Class
Attribute
Contact Customer
C
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Advantages:• Independent modifications of ontologies in the repository possible, e.g. sharing with third parties• Ontology repository may better meet user needs for ontology manipulation• Less implementation effort on meta model side
Disadvantages:• Consistency problems, e.g. in case of modifications of the ontology or the models• Different representations of models and ontologies may require extra user training• Possible time delays in interaction btw. systems
8
Using a Hybrid Approach of Integration and Referencing
Class Relationclass
Ontology Repository
Meta Model
Class
Ontology Meta Model
Contact Customer
Notify Customer
Write E-Mail
Customer reached
Customer not reached
ActivityDecision
Subsequent
Meta Model
Model
Advantages: Disadvantages:
OWL ClassProperty
Cskos:Collection
Cskos:OrderedCollection
Pdc:title
Pdefinition
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
• Coherent conceptual approach for both ontologies and domain aspects• Independent modifications of ontologies in the repository possible, e.g. sharing with third parties• Ontology repository may better meet user needs for ontology manipulation• No consistency problem on meta model side• Not all ontology concepts required on meta model side
• Consistency problems still persist btw. ontology meta model and ontology repository• Specific ontology aspects might need to be re-implemented on meta model side
The meta modeling approach has been implemented as a multi-user plattform (ADONIS®).
Implementation Platform
a multi user plattform (ADONIS ).
External interfaces
DocumentGeneration
Process Cost
Component
Transformation
Import/Export
Acquisition Modeling Analysis Simulation Evaluation
User Interaction Adonis CommunityEdition for freedownload:
http://www.adonis-community com
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Database
CORE (Modeling subsystem)
Component community.com
9
Scenarios
Scenario A: AGES Management of Clinical Trials→ Integrating Ontologies on the Meta Model Level→ Integrating Ontologies on the Meta Model Level
Scenario B: Semantic Culture Guide→ Using a Hybrid Approach of Integration and
Referencing via Protégé
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Implementation A
ActivityDecision
Subsequent
Class Relationclass
Meta Model Term Meta Model
Term
Class
Contact Customer
Notify Customer
Write E-Mail
Customer reached
Customer not reached
DecisionMeta Model
ModelMeta ModelModel
WIN/NATAallgemein
Twenn Antragunvollständig
Twenn Antragvollständig
TFachliche
Beurteilung
TNegative
Beurteilung -Antrag hat Mängel
T
keineBeseitigung der
Mängel -Fristablauf
TPositive
Beurteilung -keine Mängel
T
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Visual views on process models based on terms
10
Goal: „Make Austria’s culture events accessible through a decentralized one-stop shop by using semantic technologies.”
SCG Project Outline
Innovative Aspects:
• Creation of an adequate Modeling Method for the description of Business Models, Business Processes and Services of the organizers of culture events.
• Creation of a Culture-Ontology for the coherent description of the involved actors, the culture events and genres and the user feedback dimensions
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
• Implementation of semantic services and serviceworkflows for the support of visitors and event organizers
• Practical trial of the semantic services and workflows
Implementation B
A ti it S b t
Class Relationclass
Ontology Repository
Meta Model
Class
Ontology Meta Model
Contact Customer
Notify Customer
Write E-Mail
Customer reached
Customer not reached
ActivityDecision
Subsequent
Meta Model
Model
OWL ClassProperty
Cskos:Collection
Cskos:OrderedCollection
Pdc:title
Pdefinition
+ ADONIS Export Plugin
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
11
Outlook: www.openmodels.at
Open Model Community
Open Model Projects
Open Model Technologies
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Community Projects Technologies
SeMFIS on Open Model
• Semantic-based Modeling Framework for Information Systems (SeMFIS)
• Provision of modeling framework, technologies, and tools to support semantic information models
• Current tasks:– Provision of a web-based modeling tool based on Java
applets (AdoWeb)– Coupling of AdoWeb and Protégé on a common platform
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
– Extension of the modeling functionalities for using ontologies and conceptual models
– …
12
SeMFIS Community
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
SeMFIS Technologies: Web-based Modeller
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
13
SeMFIS Technologies: Applet Version of Protégé incl. Plugin
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at
Thank you for your attention!
http://www.dke.univie.ac.athttp://www.dke.univie.ac.at