Decision Tree Algorithm

Comp328 tutorial 1 Kai Zhang

Outline

Introduction Example Principles

Entropy Information gain

Evaluations Demo

The problem

Given a set of training cases/objects and their attribute values, try to determine the target attribute value of new examples.

Classification Prediction

Why decision tree?

Decision trees are powerful and popular tools for classification and prediction. Decision trees represent rules, which can be understood by humans and used in knowledge system such as database.

key requirements

Attribute-value description: object or case must be

Predefined classes (target values): the target

Sufficient data: enough training cases should be

A simple example

You want to guess the outcome of next week's game between the MallRats and the Chinooks.

Available knowledge / Attribute

was the game at Home or Away was the starting time 5pm, 7pm or 9pm. Did Joe play center, or forward. whether that opponent's center was tall or not. ..

Basket ball data

What we know

The game will be away, at 9pm, and that Joe will play center on offense

A classification problem Generalizing the learned rule to new examples

Definition

Decision tree is a classifier in the form of a tree structure

Decision node: specifies a test on a single attribute Leaf node: indicates the value of the target attribute Arc/edge: split of one attribute Path: a disjunction of test to make the final decision

Decision trees classify instances or examples by starting at the root of the tree and moving through it until a leaf node.

Illustration

(1) Which to start? (root)

(2) Which node to proceed?

(3) When to stop/ come to conclusion?

Random split

The tree can grow huge These trees are hard to understand. Larger trees are typically less accurate than smaller trees.

Principled Criterion

Selection of an attribute to test at each node choosing the most useful attribute for classifying examples. information gain

measures how well a given attribute separates the training examples according to their target classification This measure is used to select among the candidate attributes at each step while growing the tree

Entropy

A measure of homogeneity of the set of examples. Given a set S of positive and negative examples of some target concept (a 2-class problem), the entropy of set S relative to this binary classification is E(S) = - p(P)log2 p(P) p(N)log2 p(N)

Suppose S has 25 examples, 15 positive and 10 negatives [15+, 10-]. Then the entropy of S relative to this classification is E(S)=-(15/25) log2(15/25) - (10/25) log2 (10/25)

Some Intuitions

The entropy is 0 if the outcome is ``certain. The entropy is maximum if we have no knowledge of the system (or any outcome is equally possible).

Entropy of a 2-class problem with regard to the portion of one of the two groups

Information Gain

Information gain measures the expected reduction in entropy, or uncertainty.

vValues ( A)

Sv Entropy(Sv ) S

Values(A) is the set of all possible values for attribute A, and Sv the subset of S for which attribute A has value v Sv = {s in S | A(s) = v}. the first term in the equation for Gain is just the entropy of the original collection S the second term is the expected value of the entropy after S is partitioned using attribute A

It is simply the expected reduction in entropy caused by partitioning the examples according to this attribute. It is the number of bits saved when encoding the target value of an arbitrary member of S, by knowing the value of attribute A.

Examples

Before partitioning, the entropy is

H(10/20, 10/20) = - 10/20 log(10/20) - 10/20 log(10/20) = 1 Entropy of the first set H(home) = - 6/12 log(6/12) - 6/12 log(6/12) = 1 Entropy of the second set H(away) = - 4/8 log(6/8) - 4/8 log(4/8) = 1 12/20 * H(home) + 8/20 * H(away) = 1

Using the ``where attribute, divide into 2 subsets

Expected entropy after partitioning

Using the ``when attribute, divide into 3 subsets

Entropy of the first set H(5pm) = - 1/4 log(1/4) - 3/4 log(3/4); Entropy of the second set H(7pm) = - 9/12 log(9/12) - 3/12 log(3/12); Entropy of the second set H(9pm) = - 0/4 log(0/4) - 4/4 log(4/4) = 0

Expected entropy after partitioning

4/20 * H(1/4, 3/4) + 12/20 * H(9/12, 3/12) + 4/20 * H(0/4, 4/4) = 0.65 Information gain 1-0.65 = 0.35

Decision

Knowing the ``when attribute values provides larger information gain than ``where. Therefore the ``when attribute should be chosen for testing prior to the ``where attribute. Similarly, we can compute the information gain for other attributes. At each node, choose the attribute with the largest information gain.

Stopping rule

Every attribute has already been included along this path through the tree, or The training examples associated with this leaf node all have the same target attribute value (i.e., their entropy is zero).

Demo

Continuous Attribute?

Each non-leaf node is a test, its edge partitioning the attribute into subsets (easy for discrete attribute). For continuous attribute

Partition the continuous value of attribute A into a discrete set of intervals Create a new boolean attribute Ac , looking for a threshold c,

true if Ac c Ac false otherwise

Evaluation

Training accuracy

How many training instances can be correctly classify based on the available data? Is high when the tree is deep/large, or when there is less confliction in the training instances. however, higher training accuracy does not mean good generalization Given a number of new instances, how many of them can we correctly classify? Cross validation

Testing accuracy

Strengths

can generate understandable rules perform classification without much computation can handle continuous and categorical variables provide a clear indication of which fields are most important for prediction or classification

Weakness

Not suitable for prediction of continuous attribute. Perform poorly with many class and small data. Computationally expensive to train.

At each node, each candidate splitting field must be sorted before its best split can be found. In some algorithms, combinations of fields are used and a search must be made for optimal combining weights. Pruning algorithms can also be expensive since many candidate sub-trees must be formed and compared.

Do not treat well non-rectangular regions.