Feature Selection Based Coral Reefs Optimization for Breast Cancer Classification

Abstract

AbstractEarly detection of cancer cases is one of the most important things that help complete treatment and disappearance of this disease from the human body. Breast cancer is the most widespread invasive cancer in women, and after lung cancer, it is the second leading cause of cancer death in women. The first symptoms of breast cancer usually appear as an area of thickened tissue in the breast or a lump in the breast or an armpit. Consequently, many features can be found to indicate the existence of cancer or not. This chapter employs the coral reefs optimization (CRO) algorithm for feature selection; the CRO has shown to be very effective with various classification approaches. In this chapter, we used five standard classifiers: Logistic Regression (LR), K-nearest neighbor (KNN), Support Vector Machine with Radial Basis Function (SVM-RBF), Random Forest (RF), Decision Tree (DT). All These classifiers are presented with and without feature selection using the CRO algorithm. The results indicated that using the feature selection based on CRO achieved better results before the feature selection. The most common dataset called Breast Diagnostic Cancer Wisconsin (BDCW) is utilized to select the most significant and classify the cancer cases with a tested accuracy of 100%, 99.1%, 100%, 100%, and 100% using LR, KNN, SVM-RBF, RF, and DT respectively.