Feature Selection using Extra Trees for Breast Cancer Prediction

Abstract

Breast cancer is a disease that seriously threatens women's health. It is considering a common death cause in women. Machine learning has made significant progress in recent years to improve the effectiveness of early diagnosis of various diseases. Accurate predication and detection help decrease the death rate of breast cancer. This paper aims to predict breast cancer using several machine-learning techniques. The idea is to lower the number of features in the Wisconsin Breast Cancer Dataset (WCDB) and use it for prediction. The study used the extra trees method for feature selection and Random forest, Logistic regression, and Support Vector Machine (SVM) for testing the dataset. According to the results, SVM achieved the best performance among the other models with 98% accuracy. The proposed method in this study proved its effectiveness in breast cancer prediction.