Performance Analysis of Machine Learning Techniques for Predicting Breast Cancer
Abstract
Women of the world generally face a fatal disease problem known as breast cancer disease. Timely prediction can control the disease and save human lives. Machine learning (ML) techniques provide effective results in predicting medical problems. Therefore, a comparative analysis of four ML techniques, namely Gaussian Naive Bayes (GNB), K-Nearest Neighbors (KNN), Decision Tree (DT), and Extremely Randomized Trees / Extra Trees (ET) has been conducted in this paper to predict breast cancer at an early stage. The comparative analysis has been evaluated in terms of five performance metrics, which are Accuracy, Precision, Sensitivity, F1-Score, and Area Under the ROC Curve (AUC). The analysis results show that among the four ML techniques, Extremely Randomized Trees / Extra Trees (ET) technique gives the best results.