Practical Data Analysis Cookbook

Classifying calls with decision trees

Decision trees have been widely used to solve classification problems. A decision tree is, as the name suggests, a tree-like structure that branches out from the root. At each branch (decision) node, the remaining data is split into two groups given a decision criterion with a specified objective. The process continues until no more divisions can be made or all the samples in the ending node (a leaf) belong to the same class (that is, variance is minimized).

Getting ready

To execute this recipe, you will need pandas and Scikit-learn. To execute an alternate way of estimating a decision tree classifier, you will need mlpy. No other prerequisites are required.

How to do it…

Scikit-learn provides the DecisionTreeClassifier(...) ...