Growing a Decision Tree

Predict a target value by a sequence of questions to form or grow a decision tree. A sample here shows how to grow a decision tree.

A decision tree predicts a target value by asking a sequence of questions. At a given stage in the sequence, the question that is asked depends upon the answers to the previous questions. The goal is to ask questions that, taken together, uniquely identify specific target values. Graphically, this process forms a tree structure.

Figure 7-10 Sample Decision Tree

Description of "Figure 7-10 Sample Decision Tree"

The figure is a decision tree with nine nodes (and nine corresponding rules). The target attribute is binary: 1 if the customer increases spending, 0 if the customer does not increase spending. The first split in the tree is based on the CUST_MARITAL_STATUS attribute. The root of the tree (node 0) is split into nodes 1 and 3. Married customers are in node 1; single customers are in node 3.

The rule associated with node 1 is:

Node 1 recordCount=712,0 Count=382, 1 Count=330
CUST_MARITAL_STATUS isIN  "Married",surrogate:HOUSEHOLD_SIZE isIn "3""4-5"

Node 1 has 712 records (cases). In all 712 cases, the CUST_MARITAL_STATUS attribute indicates that the customer is married. Of these, 382 have a target of 0 (not likely to increase spending), and 330 have a target of 1 (likely to increase spending).