Breast Cancer Patient Classification from Risk Factor Analysis Using Machine Learning Classifiers

Abstract

AbstractBreast cancer-related mortality in females is a rising global health problem. The fact that for the greater duration of its course of spread, it shows no clinical symptoms, further makes it more challenging to control. Earlier diagnosis has deciding role on prognosis of disease. Innovative diagnostic techniques have provided a large database of disease. These databases, with support of machine learning, provide us a framework to arrive at a decision. This paper aims to find machine learning usefulness, its techniques and algorithms for breast cancer prediction. In this work, classifiers such as naïve Bayes, random forest, sequential minimum optimization (SMO) and logistic regression are used for classification of breast cancer patients based on different risk factors. Performance of these machine learning algorithms is analyzed based on accuracy, F-measure, recall and precision using Waikato Environment for Knowledge Analysis (WEKA). Naïve Bayes classifier has given the accurate classification of patients at high risk and low risk of breast cancer.KeywordsBreast cancerMachine learningRisk factorsAnd Risk assessmentRandom forest (RF)Support vector machine (SVM)Sequential minimum optimization (SMO)