Feature Selection for Cryotherapy and Immunotherapy Treatment Methods Based on Gravitational Search Algorithm

Abstract

Feature selection has been an active area of research which aims at identifying the most optimal subset of features that improves classification accuracy. Medical datasets contain numerous input features and require highly accurate predictions to facilitate better diagnosis. To address this problem of feature selection in medical data, an enhanced binary version of Gravitational Search Algorithm (GSA) is proposed which is based on law of gravity and attraction of masses. The proposed algorithm combines speed of Random Forest Classifier and optimization behavior of GSA. In this paper, a comprehensive review of our wrapper based proposed algorithm on Immunotherapy and Cryotherapy datasets for warts treatment using Random Forest Classifier to predict the response of the treatment is provided. The experimental results of the study show significant improvement in the accuracy of prediction.