A Hybrid Feature Selection Method based on IGSBFS and Naive Bayes for the Diagnosis of Erythemato - Squamous Diseases

Abstract

In this paper a diagnostic model based on Naive Bayes developed to diagnose erytemato squamous diseases. The hybrid feature selection method, named IGSBFS (Information Gain and Sequential Backward Floating Search), combines the advantages of filters and wrappers to select the optimal feature subset from the original feature set. In IGSBFS, Information Gain acts as filters to remove redundant features and SBFS with Naïve Bayes acts as the wrappers to select the ideal feature subset from the remaining features We conducted experiments in WEKA with 10 fold cross validation. The algorithm selected an optimum feature subset of 10 features with 98.9% accuracy.