Random Forest Weighting based Feature Selection for C4.5 Algorithm on Wart Treatment Selection Method

Abstract

Research in the field of health, especially treatment of wart disease has been widely practiced. One of the research topics related to the treatment of wart disease is in order to provide the most appropriate treatment method recommendations. Treatment methods are widely used by doctors for treatment of patients with wart disease that is the method of cryotherapy and immunotherapy. Previous research has been done on cryotherapy and immunotherapy datasets which resulted in two different prediction methods, but the accuracy level has not been satisfactory. In this study, two datasets are combined to produce a single prediction method. The method uses C4.5 algorithm combined with Random Forest Feature Weighting (C4.5+RFFW) used to select the relevant features to improve accuracy. Experimental results show that the proposed method can improve performance with accuracy and informedness are 87.22% and 71.24%, respectively. These results further facilitate physicians in determining treatment methods for patients with a single predictive method and better-predicted performance.