Information Gain Calculator

This node allows to calculate the information gain score for all features of a given dataset. The provided output is a table with all feature names in the dataset and an associated information gain score. An explanation of the information gain feature selection criterion can be found for example in “A Comparative Study on Feature Selection in Text Categorization”, Yiming Yang, Jan O. Pedersen, 1997.

The information gain measure is usually employed to select the best split in a tree node when building decision trees. This node allows to calculate the information gain values for a list of features and output it as a single list, so that the worth of a given features can be analyzed conveniently.