Comparison of Bayes Classifiers for Breast Cancer Classification

Abstract

Data analytics play vital roles in diagnosis and treatment in the health care sector. To enable practitioner decision-making, huge volumes of data should be processed with machine learning techniques to produce tools for prediction and classification. Diseases like breast cancer can be classified based on the nature of the tumor. Finding an effective algorithm for classification should help resolve the challenges present in analyzing large volume of data. The objective with this paper was to present a report on the performance of Bayes classifiers like Tree Augmented Naive Bayes (TAN), Boosted Augmented Naive Bayes (BAN) and Bayes Belief Network (BBN). Among the three approaches, TAN produced the best performance regarding classification and accuracy. The results obtained provide clear evidence for benefits of TAN usage in breast cancer classification. Applications of various machine learning algorithms could clearly assist breast cancer control efforts for identification, prediction, prevention and health care planning.