Upload
vimal-raja
View
219
Download
0
Embed Size (px)
Citation preview
8/2/2019 7 Classification
1/63
1
Data Mining:Concepts and Techniques
Classification and Prediction
8/2/2019 7 Classification
2/63
2
What is classification?
What is prediction?
8/2/2019 7 Classification
3/63
3
Classification:
predicts categorical class labels
classifies data (constructs a model) based on thetraining set and the values (class labels) in a
classifying attribute and uses it in classifying new data Prediction:
models continuous-valued functions, i.e., predictsunknown or missing values
Typical Applications credit approval
target marketing
medical diagnosis
treatment effectiveness analysis
Classification vs. Prediction
8/2/2019 7 Classification
4/63
4
ClassificationA Two-Step Process
Model construction: describing a set of predetermined classes
Each tuple/sample is assumed to belong to a predefined class,as determined by the class label attribute
The set of tuples used for model construction: training set
The model is represented as classification rules, decision trees,or mathematical formulae
Model usage: for classifying future or unknown objects
Estimate accuracy of the model
The known label of test sample is compared with the
classified result from the model Accuracy rate is the percentage of test set samples that are
correctly classified by the model
Test set is independent of training set, otherwise over-fittingwill occur
8/2/2019 7 Classification
5/63
5
Classification Process (1): ModelConstruction
Training
Data
N A M E R A N K Y E A R S T E N U R E D
Mike Assistant Prof 3 no
Mary Assistant Prof 7 yesBill Professor 2 yes
Jim Associate Prof 7 yes
Dave Assistant Prof 6 no
Anne Associate Prof 3 no
Classification
Algorithms
IF rank = professor
OR years > 6
THEN tenured = yes
Classifier
(Model)
8/2/2019 7 Classification
6/63
6
Classification Process (2): Use theModel in Prediction
Classifier
Testing
Data
N A M E R A N K Y E A R S T E N U R E D
Tom Assistant Prof 2 no
Merlisa Associate Prof 7 no
George Professor 5 yes
Joseph Assistant Prof 7 yes
Unseen Data
(Jeff, Professor, 4)
Tenured?
8/2/2019 7 Classification
7/637
Supervised vs. UnsupervisedLearning
Supervised learning (classification)
Supervision: The training data (observations,
measurements, etc.) are accompanied by labels
indicating the class of the observations
New data is classified based on the training set
Unsupervised learning (clustering)
The class labels of training data is unknown Given a set of measurements, observations, etc. with
the aim of establishing the existence of classes or
clusters in the data
8/2/2019 7 Classification
8/638
Issues regarding classification and prediction
8/2/2019 7 Classification
9/639
Issues regarding classification andprediction (1): Data Preparation
Data cleaning
Preprocess data in order to reduce noise and handle
missing values
Relevance analysis (feature selection)
Remove the irrelevant or redundant attributes
Data transformation
Generalize and/or normalize data
8/2/2019 7 Classification
10/6310
Issues regarding classification and prediction(2): Evaluating Classification Methods
Predictive accuracy (correctly predict the class of newdata)
Speed and scalability time to construct the model
time to use the model Robustness
handling noise and missing values Scalability
efficiency in disk-resident databases Interpretability:
understanding and insight provided by the model Goodness of rules
decision tree size compactness of classification rules
8/2/2019 7 Classification
11/6311
Classification by decision tree induction
8/2/2019 7 Classification
12/6312
Classification by Decision TreeInduction
Decision tree
A flow-chart-like tree structure
Internal node denotes a test on an attribute
Branch represents an outcome of the test
Leaf nodes represent class labels or class distribution Decision tree generation consists of two phases
Tree construction
At start, all the training examples are at the root
Partition examples recursively based on selected attributes
Tree pruning
Identify and remove branches that reflect noise or outliers
Use of decision tree: Classifying an unknown sample
Test the attribute values of the sample against the decision tree
8/2/2019 7 Classification
13/6313
Training Dataset
age income student credit_rating
40 low yes fair
>40 low yes excellent
3140 low yes excellent
8/2/2019 7 Classification
14/6314
Output: A Decision Tree for buys_computer
age?
overcast
student? credit rating?
no yes fairexcellent
40
no noyes yes
yes
30..40
8/2/2019 7 Classification
15/6315
Algorithm for Decision Tree Induction
Basic algorithm (a greedy algorithm)
Tree is constructed in a top-down recursive divide-and-conquermanner
At start, all the training examples are at the root
Attributes are categorical (if continuous-valued, they arediscretized in advance)
Examples are partitioned recursively based on selected attributes
Test attributes are selected on the basis of a heuristic orstatistical measure (e.g., information gain)
Conditions for stopping partitioning
All samples for a given node belong to the same class
There are no remaining attributes for further partitioningmajority voting is employed for classifying the leaf
There are no samples left
8/2/2019 7 Classification
16/6316
Attribute Selection Measure
Information gain
All attributes are assumed to be categorical Can be modified for continuous-valued attributes
8/2/2019 7 Classification
17/6317
Information Gain
Select the attribute with the highest information gain
Assume there are two classes, P and N
Let the set of examples Scontain pelements of class P
and nelements of class N
The amount of information, needed to decide if an
arbitrary example in Sbelongs to P or Nis defined as
np
n
np
n
np
p
np
pnpI
22 loglog),(
8/2/2019 7 Classification
18/6318
Information Gain in DecisionTree Induction
Assume that using attribute A a set Swill be partitioned
into sets {S1, S2, , Sv}
IfSicontains piexamples ofPand niexamples ofN,
the entropy, or the expected information needed toclassify objects in all subtrees Si is
The encoding information that would be gained by
branching onA
1
),()(
i
iiii npI
np
npAE
)(),()( AEnpIAGain
8/2/2019 7 Classification
19/6319
Attribute Selection by InformationGain Computation
Class P: buys_computer =
yes
Class N: buys_computer =
no
I(p, n) = I(9, 5) =0.940
Compute the entropy for age:
Hence
Similarly
age pi ni I(pi, ni)40 3 2 0.971
69.0)2,3(14
5
)0,4(14
4)3,2(
14
5)(
I
IIageE
048.0)_(
151.0)(
029.0)(
ratingcreditGain
studentGain
incomeGain
)(),()( ageEnpIageGain
8/2/2019 7 Classification
20/6320
Extracting Classification Rules from Trees
Represent the knowledge in the form ofIF-THEN rules
One rule is created for each path from the root to a leaf
Each attribute-value pair along a path forms a conjunction
The leaf node holds the class prediction
Rules are easier for humans to understand
Example
IF age= 40 AND credit_rating= fair THEN buys_computer=no
8/2/2019 7 Classification
21/6321
Avoid Overfitting in Classification
The generated tree may overfit the training data
Too many branches, some may reflect anomaliesdue to noise or outliers
Result is in poor accuracy for unseen samples
Two approaches to avoid overfitting Prepruning: Halt tree construction earlydo not split
a node if this would result in the goodness measurefalling below a threshold
Difficult to choose an appropriate threshold Postpruning: Remove branches from a fully grown
treeget a sequence of progressively pruned trees
Use a set of data different from the training datato decide which is the best pruned tree
8/2/2019 7 Classification
22/6322
Approaches to Determine the FinalTree Size
Separate training (2/3) and testing (1/3) sets
Use cross validation, e.g., 10-fold cross validation
Use all the data for training but apply a statistical test (e.g., chi-square) to
estimate whether expanding or pruning a
node may improve the entire distribution
Use minimum description length (MDL) principle:
halting growth of the tree when the encoding
is minimized
8/2/2019 7 Classification
23/63
23
Enhancements to basic decisiontree induction
Allow for continuous-valued attributes
Dynamically define new discrete-valued attributes thatpartition the continuous attribute value into a discreteset of intervals
Handle missing attribute values
Assign the most common value of the attribute
Assign probability to each of the possible values
Attribute construction
Create new attributes based on existing ones that aresparsely represented
This reduces fragmentation, repetition, and replication
8/2/2019 7 Classification
24/63
24
Classification in Large Databases
Classificationa classical problem extensively studied by
statisticians and machine learning researchers
Scalability: Classifying data sets with millions of examples
and hundreds of attributes with reasonable speed Why decision tree induction in data mining?
relatively faster learning speed (than other classificationmethods)
convertible to simple and easy to understandclassification rules
can use SQL queries for accessing databases
comparable classification accuracy with other methods
8/2/2019 7 Classification
25/63
25
Scalable Decision Tree InductionMethods in Data Mining Studies
SLIQ
builds an index for each attribute and only class list andthe current attribute list reside in memory
SPRINT
constructs an attribute list data structure
PUBLIC
integrates tree splitting and tree pruning: stop growing
the tree earlier RainForest
separates the scalability aspects from the criteria thatdetermine the quality of the tree
builds an AVC-list (attribute, value, class label)
8/2/2019 7 Classification
26/63
26
Data Cube-Based Decision-TreeInduction
Integration of generalization with decision-tree induction
Classification at primitive concept levels
E.g., precise temperature, humidity, outlook, etc.
Low-level concepts, scattered classes, bushyclassification-trees
Semantic interpretation problems.
Cube-based multi-level classification Relevance analysis at multi-levels.
Information-gain analysis with dimension + level.
8/2/2019 7 Classification
27/63
27
Presentation of Classification Results
8/2/2019 7 Classification
28/63
28
Bayesian Classification
8/2/2019 7 Classification
29/63
29
Bayesian Classification: Why?
Probabilistic learning: Calculate explicit probabilities forhypothesis, among the most practical approaches to certaintypes of learning problems
Incremental: Each training example can incrementally
increase/decrease the probability that a hypothesis iscorrect. Prior knowledge can be combined with observeddata.
Probabilistic prediction: Predict multiple hypotheses,
weighted by their probabilities Standard: Even when Bayesian methods are
computationally intractable, they can provide a standard ofoptimal decision making against which other methods can
be measured
8/2/2019 7 Classification
30/63
30
Bayesian Theorem
Given training data D, posteriori probability of ahypothesis h, P(h|D)follows the Bayes theorem
Practical difficulty: require initial knowledge of manyprobabilities, significant computational cost
)(
)()|()|(
DP
hPhDPDhP
8/2/2019 7 Classification
31/63
31
Naive Bayesian Classifier (simple Bayesianclassifier)works as
Each data sample is represented by an n-dimensionalfeature vector, X=(x1, x2,xn), depicting nmeasurements on the sample from n attributes,
respectively, A1,A2,,An. Suppose there are m classes, C1,C2,Cm. X is data
sample (ie having no class label), the classifier willpredict that X belongs to the class having the highest
posterior probability, conditioned on X. ie the naveBayesian classifier assigns an unknown sample X toclass Ci iff
P(Ci/X) = P(X/Ci)P(Ci) / P(X)
8/2/2019 7 Classification
32/63
32
Naive Bayesian Classifier works as
P(X) is constant for all classes, only P(X/Ci)P(Ci) need to bemaximized. Class prior probabilities may be estimated by P(Ci)=si/s , where si is the no. of training samples of class Ci, and s is the totalno. of training samples.
If we have data sets with many attributes, it would be
computationally expensive to compute P(X/Ci). To reducecomputation in evaluating P(X/Ci), the nave assumption of classconditional independence is made. This presumes that the valuesof the attributes are conditionally independent of one another,given the class label of the sample ie there are no dependence
relationships among the attributes. Thus,n
P(X/Ci)=p P(xk/ Ci)
k=1
8/2/2019 7 Classification
33/63
33
Naive Bayesian Classifier works as
the probabilities P(x1/Ci), P(x2/Ci), , P(xn/Ci) can be estimatedfrom training samples, where
- If Ak is categorical, then P(xk/Ci)= sik/si,
where sikis the no. of training samples of class Ci having the
value xk for Akand si is the no. of training samples belonging toCi.
- If Ak is continuous- valued, then the attribute is assumed tohave a Gaussian distribution so that
P(xk/C
i) = g(x
k,m
Ci,s
Ci) =
where g(xk,mCi ,s Ci) is the Gaussian (normal) density functionfor attribute Ak, while mCi and s Ci are the mean and standarddeviation respectively, given the values for attribute for Aktraining samples of class Ci.
8/2/2019 7 Classification
34/63
34
Naive Bayesian Classifier works as
In order to classify an unknown sample X, P(X/Ci)P(Ci)is evaluated for each class Ci. Sample X is thenassigned to the class Ci iff
P(X/Ci)P(Ci) > P(X/Cj)P(Cj)
for 1
8/2/2019 7 Classification
35/63
35
Bayesian classification
The classification problem may be formalizedusing a-posteriori probabilities:
P(C|X) = prob. that the sample tuple
X= is of class C.
E.g. P(class=N | outlook=sunny,windy=true,)
Idea: assign to sampleXthe class labelCsuchthatP(C|X) is maximal
8/2/2019 7 Classification
36/63
36
Estimating a-posteriori probabilities
Bayes theorem:
P(C|X) = P(X|C)P(C) / P(X)
P(X) is constant for all classes
P(C) = relative freq of class C samples
C such that P(C|X) is maximum =C such that P(X|C)P(C) is maximum
Problem: computing P(X|C) is unfeasible!
8/2/2019 7 Classification
37/63
37
The independence hypothesis
makes computation possible
yields optimal classifiers when satisfied
but is seldom satisfied in practice, as attributes
(variables) are often correlated.
Attempts to overcome this limitation:
Bayesian networks, that combine Bayesian reasoning
with causal relationships between attributes
Decision trees, that reason on one attribute at the
time, considering most important attributes first
8/2/2019 7 Classification
38/63
38
Bayesian Belief Networks
FamilyHistory
LungCancer
PositiveXRay
Smoker
Emphysema
Dyspnea
LC
~LC
(FH, S) (FH, ~S)(~FH, S)(~FH, ~S)
0.8
0.2
0.5
0.5
0.7
0.3
0.1
0.9
Bayesian Belief Networks
The conditional probability table
for the variable LungCancer
8/2/2019 7 Classification
39/63
39
Bayesian Belief Networks
Bayesian belief network allows a subsetof the variables
conditionally independent
A graphical model of causal relationships
Several cases of learning Bayesian belief networks
Given both network structure and all the variables:
easy
Given network structure but only some variables
When the network structure is not known in advance
8/2/2019 7 Classification
40/63
40
Classification by backpropagation
8/2/2019 7 Classification
41/63
41
Neural Networks
Advantages
prediction accuracy is generally high
robust, works when training examples contain errors
output may be discrete, real-valued, or a vector ofseveral discrete or real-valued attributes
fast evaluation of the learned target function
Criticism
long training time difficult to understand the learned function (weights)
not easy to incorporate domain knowledge
8/2/2019 7 Classification
42/63
42
A Neuron
The n-dimensional input vector xis mapped intovariable yby means of the scalar product and anonlinear function mapping
mk-
f
weighted
sum
Input
vectorx
outputy
Activation
function
weight
vector w
w0
w1
wn
x0
x1
xn
8/2/2019 7 Classification
43/63
43
Network Training
The ultimate objective of training
obtain a set of weights that makes almost all the
tuples in the training data classified correctly
Steps Initialize weights with random values
Feed the input tuples into the network one by one
For each unit
Compute the net input to the unit as a linear combinationof all the inputs to the unit
Compute the output value using the activation function
Compute the error
Update the weights and the bias
8/2/2019 7 Classification
44/63
44
Multi-Layer Perceptron
Output nodes
Input nodes
Hidden nodes
Output vector
Input vector:xi
wij
i
jiijj OwI
jIj eO
1
1
))(1( jjjjj OTOOErr
jkk
kjjj wErrOOErr )1(
ijijij OErrlww )(jjj
Errl)(
8/2/2019 7 Classification
45/63
45
Classification based on concepts from
association rule mining
8/2/2019 7 Classification
46/63
46
Association-Based Classification
Several methods for association-based classification
ARCS: Quantitative association mining and clusteringof association rules
It beats C4.5 in (mainly) scalability and also accuracy
Associative classification:
It mines high support and high confidence rules in the form ofcond_set => y, where y is a class label
CAEP (Classification by aggregating emerging patterns)
Emerging patterns (EPs): the itemsets whose supportincreases significantly from one class to another
Mine Eps based on minimum support and growth rate
8/2/2019 7 Classification
47/63
47
Other Classification Methods
8/2/2019 7 Classification
48/63
48
Other Classification Methods
k-nearest neighbor classifier
case-based reasoning
Genetic algorithm Rough set approach
Fuzzy set approaches
8/2/2019 7 Classification
49/63
49
Instance-Based Methods
Instance-based learning: Store training examples and delay the processing
(lazy evaluation) until a new instance must beclassified
Typical approaches k-nearest neighbor approach
Instances represented as points in a Euclideanspace.
Locally weighted regression
Constructs local approximation
Case-based reasoning
Uses symbolic representations and knowledge-based inference
8/2/2019 7 Classification
50/63
50
The k-Nearest Neighbor Algorithm
All instances correspond to points in the n-D space. The nearest neighbor are defined in terms of
Euclidean distance.
The target function could be discrete- or real- valued.
For discrete-valued, the k-NN returns the mostcommon value among the k training examples nearesttoxq.
Vonoroi diagram: the decision surface induced by 1-NN for a typical set of training examples.
.
_+
_ xq
+
_ _+
_
_
+
.
.
.
. .
8/2/2019 7 Classification
51/63
51
Discussion on the k-NN Algorithm
The k-NN algorithm for continuous-valued target functions Calculate the mean values of the knearest neighbors
Distance-weighted nearest neighbor algorithm
Weight the contribution of each of the k neighbors
according to their distance to the query point xq
giving greater weight to closer neighbors
Similarly, for real-valued target functions
Robust to noisy data by averaging k-nearest neighbors
Curse of dimensionality: distance between neighbors couldbe dominated by irrelevant attributes.
To overcome it, axes stretch or elimination of the leastrelevant attributes.
wd xq xi
12( , )
8/2/2019 7 Classification
52/63
52
Case-Based Reasoning
Also uses: lazy evaluation + analyze similar instances Difference: Instances are not points in a Euclidean space
Example: Water faucet problem in CADET
Methodology
Instances represented by rich symbolic descriptions(e.g., function graphs)
Multiple retrieved cases may be combined
Tight coupling between case retrieval, knowledge-based
reasoning, and problem solving Research issues
Indexing based on syntactic similarity measure, andwhen failure, backtracking, and adapting to additionalcases
8/2/2019 7 Classification
53/63
53
Remarks on Lazy vs. Eager Learning
Instance-based learning: lazy evaluation Decision-tree and Bayesian classification: eager evaluation
Key differences
Lazy method may consider query instance xqwhen deciding how togeneralize beyond the training data D
Eager method cannot since they have already chosen globalapproximation when seeing the query
Efficiency: Lazy - less time training but more time predicting
Accuracy
Lazy method effectively uses a richer hypothesis space since it usesmany local linear functions to form its implicit global approximationto the target function
Eager: must commit to a single hypothesis that covers the entireinstance space
8/2/2019 7 Classification
54/63
54
Genetic Algorithms
GA: based on an analogy to biological evolution Each rule is represented by a string of bits
An initial population is created consisting of randomlygenerated rules
e.g., IF A1 and Not A2 then C2 can be encoded as 100
Based on the notion of survival of the fittest, a newpopulation is formed to consists of the fittest rules andtheir offsprings
The fitness of a rule is represented by its classificationaccuracy on a set of training examples
Offsprings are generated by crossover and mutation
8/2/2019 7 Classification
55/63
55
Rough Set Approach
Rough sets are used to approximately or roughlydefine equivalent classes
A rough set for a given class C is approximated by twosets: a lower approximation (certain to be in C) and an
upper approximation (cannot be described as notbelonging to C)
Finding the minimal subsets (reducts) of attributes (forfeature reduction) is NP-hard but a discernibility matrix
is used to reduce the computation intensity
Fuzzy Set
8/2/2019 7 Classification
56/63
56
Fuzzy SetApproaches
Fuzzy logic uses truth values between 0.0 and 1.0 torepresent the degree of membership (such as usingfuzzy membership graph)
Attribute values are converted to fuzzy values e.g., income is mapped into the discrete categories
{low, medium, high} with fuzzy values calculated
For a given new sample, more than one fuzzy value may
apply Each applicable rule contributes a vote for membership
in the categories
Typically, the truth values for each predicted categoryare summed
8/2/2019 7 Classification
57/63
57
Prediction
8/2/2019 7 Classification
58/63
58
What Is Prediction?
Prediction is similar to classification
First, construct a model
Second, use model to predict unknown value
Major method for prediction is regression
Linear and multiple regression
Non-linear regression
Prediction is different from classification Classification refers to predict categorical class label
Prediction models continuous-valued functions
8/2/2019 7 Classification
59/63
59
Predictive modeling: Predict data values or constructgeneralized linear models based on the database data.
One can only predict value ranges or category distributions
Method outline:
Minimal generalization Attribute relevance analysis
Generalized linear model construction
Prediction
Determine the major factors which influence the prediction Data relevance analysis: uncertainty measurement,
entropy analysis, expert judgement, etc.
Multi-level prediction: drill-down and roll-up analysis
Predictive Modeling in Databases
Regress Analysis and Log-Linear
8/2/2019 7 Classification
60/63
60
Linear regression: Y = + X Two parameters , and specify the line and are to
be estimated by using the data at hand.
using the least squares criterion to the known values
of Y1, Y2, , X1, X2, . Multiple regression: Y = b0 + b1 X1 + b2 X2.
Many nonlinear functions can be transformed into theabove.
Log-linear models:
The multi-way table of joint probabilities isapproximated by a product of lower-order tables.
Probability: p(a, b, c, d) =abacadbcd
Regress Analysis and Log-LinearModels in Prediction
8/2/2019 7 Classification
61/63
61
Locally Weighted Regression
Construct an explicit approximation to fover a local regionsurrounding query instance xq.
Locally weighted linear regression:
The target function fis approximated near xqusing thelinear function:
minimize the squared error: distance-decreasing weightK
the gradient descent training rule:
In most cases, the target function is approximated by a
constant, linear, or quadratic function.
( ) ( ) ( ) f x w w a x wnan x 0 1 1
E xq f x f x x k nearest neighbors of xq
K d xq x( ) ( ( )( ))
_ _ _ _( ( , ))
12
2
wj
K d xq x f x f x ajx
x k nearest neighbors of xq
( ( , ))(( ( ) ( )) ( )_ _ _ _
8/2/2019 7 Classification
62/63
62
Prediction: Numerical Data
P di ti C t i l D t
8/2/2019 7 Classification
63/63
Prediction: Categorical Data