Upload
rajendra-akerkar
View
6.657
Download
0
Tags:
Embed Size (px)
Citation preview
Decision TreeDecision Tree
R. AkerkarTMRF, Kolhapur, India
1R. Akerkar
Introduction
A classification scheme which generates a tree and ga set of rules from given data set.
Th t f d il bl f d l i The set of records available for developing classification methods is divided into two disjoint subsets – a training set and a test set.g
The attributes of the records are categorise into two types:
Attrib tes hose domain is n merical are called n merical Attributes whose domain is numerical are called numerical attributes.
Attributes whose domain is not numerical are called the categorical attributescategorical attributes.
2R. Akerkar
Introduction
A decision tree is a tree with the following properties:g p p An inner node represents an attribute. An edge represents a test on the attribute of the father
nodenode. A leaf represents one of the classes.
Construction of a decision tree Based on the training data
Top Down strategy Top-Down strategy
3R. Akerkar
Decision TreeExample
The data set has five attributes. There is a special attribute: the attribute class is the class label. The attributes, temp (temperature) and humidity are numerical
attributes Other attributes are categorical, that is, they cannot be ordered.
Based on the training data set, we want to find a set of rules to know what values of outlook, temperature, humidity and wind, determine whether or not to play golf.determine whether or not to play golf.
4R. Akerkar
Decision TreeExample
We have five leaf nodes. In a decision tree, each leaf node represents a rule.
We have the following rules corresponding to the tree given in Figure.
RULE 1 If it is sunny and the humidity is not above 75% then play RULE 1 If it is sunny and the humidity is not above 75%, then play. RULE 2 If it is sunny and the humidity is above 75%, then do not play. RULE 3 If it is overcast, then play. RULE 4 If it is rainy and not windy, then play.
RULE 5 If it i i d i d th d 't l RULE 5 If it is rainy and windy, then don't play.
5R. Akerkar
Classification
The classification of an unknown input vector is done by p ytraversing the tree from the root node to a leaf node.
A record enters the tree at the root node. At the root a test is applied to determine which child At the root, a test is applied to determine which child
node the record will encounter next. This process is repeated until the record arrives at a leaf
dnode. All the records that end up at a given leaf of the tree are
classified in the same way. y There is a unique path from the root to each leaf. The path is a rule which is used to classify the records.
6R. Akerkar
In our tree, we can carry out the classification for an unknown record as follows. o a u o eco d as o o s
Let us assume, for the record, that we know the values of the first four attributes (but wethe values of the first four attributes (but we do not know the value of class attribute) as
outlook= rain; temp = 70; humidity = 65; and windy= truewindy true.
7R. Akerkar
We start from the root node to check the value of the attribute associated at the root node.
This attribute is the splitting attribute at this node. For a decision tree, at every node there is an attribute associated
with the node called the splitting attribute.
In our example, outlook is the splitting attribute at root. Since for the given record, outlook = rain, we move to the right-
most child node of the root. At this node, the splitting attribute is windy and we find that for
the record we want classify, windy = true. Hence, we move to the left child node to conclude that the class
l b l I " l "label Is "no play".
8R. Akerkar
The accuracy of the classifier is determined by the percentage of the t t d t t th t i tl l ifi dtest data set that is correctly classified.
We can see that for Rule 1 there are two records of the test data set satisfying outlook= sunny and humidity < 75, and only one of these is correctly classified as play.
Thus, the accuracy of this rule is 0.5 (or 50%). Similarly, the accuracy of Rule 2 is also 0.5 (or 50%). The accuracy of Rule 3 is 0.66.
RULE 1If it is sunny and the humidity is not above 75%, then play.
9R. Akerkar
Concept of Categorical Attributes
Consider the following training data set.
There are three attributes, namely, age, pincode and class.
The attribute class is used for
class label.
The attribute age is a numeric attribute, whereas pincode is a categorical one.
Th h th d i f i d i i d i b d fi dThough the domain of pincode is numeric, no ordering can be defined among pincode values.
You cannot derive any useful information if one pin-code is greater than another pincodeanother pincode.
10R. Akerkar
Figure gives a decision tree for the training datatraining data.
The splitting attribute at the root is pincode and the splitting criterionpincode and the splitting criterion here is pincode = 500 046.
Similarly, for the left child node, the splitting criterion is age < 48 (the p g g (splitting attribute is age).
Although the right child node has At root level, we have 9 records. The associated splitting criterion is
the same attribute as the splitting attribute, the splitting criterion is different.
p gpincode = 500 046.
As a result, we split the records into two subsets. Records 1, 2, 4, 8, and 9 are to the left child note and remaining to the right node.
The process is repeated at every node.
11R. Akerkar
Advantages and Shortcomings of Decision Tree Classifications A decision tree construction process is concerned with identifying
h li i ib d li i i i l l f hthe splitting attributes and splitting criterion at every level of the tree.
Major strengths are: Decision tree able to generate understandable rules. They are able to handle both numerical and categorical attributes. They provide clear indication of which fields are most important for
prediction or classificationprediction or classification.
Weaknesses are: The process of growing a decision tree is computationally expensive At The process of growing a decision tree is computationally expensive. At
each node, each candidate splitting field is examined before its best split can be found.
Some decision tree can only deal with binary-valued target classes.
12R. Akerkar
Iterative Dichotomizer (ID3)
Quinlan (1986) Each node corresponds to a splitting attribute Each arc is a possible value of that attribute.
At each node the splitting attribute is selected to be the most informative among the attributes not yet considered in the path from the root.
Entropy is used to measure how informative is a node. The algorithm uses the criterion of information gain to determine the
d f litgoodness of a split. The attribute with the greatest information gain is taken as
the splitting attribute, and the data set is split for all distinct values of the attributevalues of the attribute.
13R. Akerkar
Training DatasetThis follows an example from Quinlan’s ID3
age income student credit_rating buys_computer<=30 high no fair no
The class label attribute, buys_computer, has two distinct values.
g<=30 high no excellent no31…40 high no fair yes>40 medium no fair yes>40 low yes fair yes
Thus there are two distinct classes. (m =2)
Class C1 corresponds to yesand class C2 corresponds to no
>40 low yes excellent no31…40 low yes excellent yes<=30 medium no fair no<=30 low yes fair yes
40 di f i
and class C2 corresponds to no.
There are 9 samples of class yesand 5 samples of class no.
>40 medium yes fair yes<=30 medium yes excellent yes31…40 medium no excellent yes31…40 high yes fair yes>40 medium no excellent no>40 medium no excellent no
14R. Akerkar
Extracting Classification Rules from Treesg
Represent the knowledge in the form of IF-THEN rules
One rule is created for each path from the root to a leafpath from the root to a leaf
Each attribute-value pair along a path forms a conjunction
The leaf node holds the class predictionp
Rules are easier for humans to understand
What are the rules?
15R. Akerkar
Solution (Rules)
IF age = “<=30” AND student = “no” THEN buys_computer = “no”
IF age = “<=30” AND student = “yes” THEN buys_computer = “yes”
IF age = “31…40” THEN buys_computer = “yes”
IF age = “>40” AND credit_rating = “excellent” THEN buys_computer = “yes”
IF age = “<=30” AND credit_rating = “fair” THEN buys_computer = “no”
16R. Akerkar
Algorithm for Decision Tree Induction
Basic algorithm (a greedy algorithm) Tree is constructed in a top-down recursive divide-and-conquer
manner At start, all the training examples are at the root Attributes are categorical (if continuous-valued, they are g ( y
discretized in advance) Examples are partitioned recursively based on selected attributes Test attributes are selected on the basis of a heuristic or
statistical measure (e.g., information gain) Conditions for stopping partitioning
All samples for a given node belong to the same classp g g There are no remaining attributes for further partitioning –
majority voting is employed for classifying the leaf There are no samples leftp
17R. Akerkar
Attribute Selection Measure: Information Gain (ID3/C4 5)Gain (ID3/C4.5)
Select the attribute with the highest information gain Select the attribute with the highest information gain S contains si tuples of class Ci for i = {1, …, m} information measures info required to classify any q y y
arbitrary tuple ….information is encoded in bits.
f b h l { }sslog
ss),...,s,ssI( i
m
i
im21 2
1
entropy of attribute A with values {a1,a2,…,av}
)s,...,s(Is
s...sE(A) mjj
v
j
mjj1
1
1
information gained by branching on attribute A
sj 1
E(A))s,...,s,I(sGain(A) m 21
18R. Akerkar
Entropy Entropy measures the homogeneity (purity) of a set of examples. It gives the information content of the set in terms of the class labels of
the examples. Consider that you have a set of examples, S with two classes, P and N. Let
the set have p instances for the class P and n instances for the class N. So the total number of instances we have is t = p + n. The view [p, n] can
be seen as a class distribution of S.
The entropy for S is defined as Entropy(S) = - (p/t).log2(p/t) - (n/t).log2(n/t)
Example: Let a set of examples consists of 9 instances for class positive, and 5 instances for class negative.
Answer: p = 9 and n = 5. So Entropy(S) = - (9/14).log2(9/14) - (5/14).log2(5/14) = -(0.64286)(-0.6375) - (0.35714)(-1.48557) = (0.40982) + (0.53056) = 0.940
19R. Akerkar
Entropy
The entropy for a completely pure set is 0 and is 1 for a set with l f b th th l
y
equal occurrences for both the classes.
i.e. Entropy[14,0] = - (14/14).log2(14/14) - (0/14).log2(0/14)1 l 2(1) 0 l 2(0)= -1.log2(1) - 0.log2(0)
= -1.0 - 0 = 0
i.e. Entropy[7,7] = - (7/14).log2(7/14) - (7/14).log2(7/14)= - (0.5).log2(0.5) - (0.5).log2(0.5)= - (0.5).(-1) - (0.5).(-1)= 0.5 + 0.5 = 1
20R. Akerkar
Attribute Selection by Information Gain ComputationComputation
Class P: buys_computer = “yes” Class N: buys_computer = “no” 5
)0,4(144)3,2(
145)( IIageE
I(p, n) = I(9, 5) =0.940 Compute the entropy for age:
“ 30” h 5age pi ni I(pi, ni)
694.0)2,3(145
I
5 means “age <=30” has 5 out of 14 samples, with 2 yes's and 3 no’s. Hence
<=30 2 3 0.97130…40 4 0 0>40 3 2 0.971
)3,2(145 I
and 3 no s. Hence
Similarly
40 3 2 0.971
029.0)( incomeGain
246.0)(),()( ageEnpIageGainage income student credit_rating buys_computer
<=30 high no fair no<=30 high no excellent no31…40 high no fair yes>40 medium no fair yes Similarly,
048.0)_(151.0)(029.0)(
ratingcreditGainstudentGainincomeGain
>40 low yes fair yes>40 low yes excellent no31…40 low yes excellent yes<=30 medium no fair no<=30 low yes fair yes>40 medium yes fair yes>40 medium yes fair yes<=30 medium yes excellent yes31…40 medium no excellent yes31…40 high yes fair yes>40 medium no excellent no
Since, age has the highest information gain among the attributes, it is selected as the test attribute. 21R. Akerkar
Exercise 1
The following table consists of training data from an employee databasedatabase.
Let status be the class attribute. Use the ID3 algorithm to construct a
decision tree from the given data.
22R. Akerkar
Solution 1
23R. Akerkar
Other Attribute Selection Measures
Gini index (CART IBM IntelligentMiner) Gini index (CART, IBM IntelligentMiner) All attributes are assumed continuous-valued Assume there exist several possible split values for each Assume there exist several possible split values for each
attribute May need other tools, such as clustering, to get the
possible split values Can be modified for categorical attributes
24R. Akerkar
Gini Index (IBM IntelligentMiner)
If a data set T contains examples from n classes, gini index, gini(T) is defined as n
Ti i 21)(defined as
where pj is the relative frequency of class j in T. If a data set T is split into two subsets T1 and T2 with sizes N1 and N2
j
p jTgini1
21)(
If a data set T is split into two subsets T1 and T2 with sizes N1 and N2respectively, the gini index of the split data contains examples from nclasses, the gini index gini(T) is defined as
)()()( 22
11 Tgini
NNTgini
NNTgini split
The attribute provides the smallest ginisplit(T) is chosen to split the node (need to enumerate all possible splitting points for each attribute).
25R. Akerkar
Exercise 2
26R. Akerkar
Solution 2 SPLIT: Age <= 50 ---------------------- | High | Low | Total -------------------- S1 (left) | 8 | 11 | 19 S2 (right) | 11 | 10 | 21 -------------------- For S1: P(high) = 8/19 = 0.42 and P(low) = 11/19 = 0.58 For S2: P(high) = 11/21 = 0.52 and P(low) = 10/21 = 0.48 For S2: P(high) 11/21 0.52 and P(low) 10/21 0.48 Gini(S1) = 1-[0.42x0.42 + 0.58x0.58] = 1-[0.18+0.34] = 1-0.52 = 0.48 Gini(S2) = 1-[0.52x0.52 + 0.48x0.48] = 1-[0.27+0.23] = 1-0.5 = 0.5 Gini-Split(Age<=50) = 19/40 x 0.48 + 21/40 x 0.5 = 0.23 + 0.26 = 0.49
SPLIT: Salary <= 65K SPLIT: Salary < 65K ---------------------- | High | Low | Total -------------------- S1 (top) | 18 | 5 | 23 S2 (bottom) | 1 | 16 | 17 S2 (bottom) | 1 | 16 | 17 -------------------- For S1: P(high) = 18/23 = 0.78 and P(low) = 5/23 = 0.22 For S2: P(high) = 1/17 = 0.06 and P(low) = 16/17 = 0.94 Gini(S1) = 1-[0.78x0.78 + 0.22x0.22] = 1-[0.61+0.05] = 1-0.66 = 0.34 Gini(S2) = 1-[0 06x0 06 + 0 94x0 94] = 1-[0 004+0 884] = 1-0 89 = 0 11 Gini(S2) = 1-[0.06x0.06 + 0.94x0.94] = 1-[0.004+0.884] = 1-0.89 = 0.11 Gini-Split(Age<=50) = 23/40 x 0.34 + 17/40 x 0.11 = 0.20 + 0.05 = 0.25
27R. Akerkar
Exercise 3
In previous exercise which is a better split of In previous exercise, which is a better split of the data among the two split points? Why?
28R. Akerkar
Solution 3
Intuitively Salary <= 65K is a better split point since it produces relatively ``pure'' partitions as opposed to Age <= 50 whichrelatively pure'' partitions as opposed to Age <= 50, which results in more mixed partitions (i.e., just look at the distribution of Highs and Lows in S1 and S2).
More formally, let us consider the properties of the Gini index. If a partition is totally pure, i.e., has all elements from the same class, then gini(S) = 1-[1x1+0x0] = 1-1 = 0 (for two classes).
On the other hand if the classes are totally mixed, i.e., both classes have equal probability then gini(S) = 1 [0 5x0 5 + 0 5x0 5] = 1 [0 25+0 25] = 0 5gini(S) = 1 - [0.5x0.5 + 0.5x0.5] = 1-[0.25+0.25] = 0.5.
In other words the closer the gini value is to 0, the better the partition is Since Salary has lower gini it is a better splitpartition is. Since Salary has lower gini it is a better split.
29R. Akerkar
Avoid Overfitting in Classificationvo d v g C ss c o
Overfitting: An induced tree may overfit the training dataOverfitting: An induced tree may overfit the training data Too many branches, some may reflect anomalies due to noise
or outliers Poor accuracy for unseen samplesy p
Two approaches to avoid overfitting Prepruning: Halt tree construction early—do not split a node if
this would result in the goodness measure falling below a threshold Difficult to choose an appropriate threshold
Postpruning: Remove branches from a “fully grown” tree—get a f i l d tsequence of progressively pruned trees
Use a set of data different from the training data to decide which is the “best pruned tree”
30R. Akerkar