Using Information Gain Attribute Evaluation to Classify Sonar Targets

Abstract

This paper presents an application of Information Gain (IG) attribute evaluation to the classification of the sonar targets with C4.5 decision tree. C4.5 decision tree has inherited ability to focus on relevant features and ignore irrelevant ones, but such method may also benefit from independent feature selection. In our experiments, IG attribute evaluation significantly improves C4.5 decision tree. This research also shows that feature selection helps increase computational efficiency while improving classification accuracy.