Feature Selection using Random Forest Classifier for Predicting Prostate Cancer

Abstract Prostate cancer is cancer that attacks the prostate gland, usually affecting men over 50 years. Prostate cancer is a disease that develops slowly. Based on this, rapid and precise detection is needed so that the disease can be treated immediately. This study focuses on the application Feature Selection using the Random Forest Classifier to detect prostate cancer. The Random Forest Classifier is a method of classifying data by determining the decision tree. The use of more trees will affect the accuracy to be obtained for the better. The Random Forest Classifier can classify data that has incomplete attributes and can be used to handle large sample data. Selection of features is an important process because it can affect the accuracy of classification. This method increases accuracy by about 87%. Thus, the selection of features can improve accuracy in the detection of prostate cancer.