Performance Assessment of Different Machine Learning Algorithms for Medical Decision Support Systems

Abstract

Recently, the significance of decision support system (DSS) is increased in diverse domains ranging from medicine to healthcare, finance to marketing. Medical DSS (MDSS) is a hot research topic that assists the physicians to choose a better method for treating the patients. MDSS involves intelligent software systems that offer decision to the physicians. In present days, machine learning algorithms finds useful to predict the data by analyzing the past behavior. This research work has made an attempt to analyze the performance of three Machine Learning (ML) models namely radial basis function (RBF), logistic regression (LR) and Naïve bayes (NB) models. The results are validated using a benchmark medical dataset and the outcome pointed out that the LR classifier model is a better option for DSS. The detailed simulation outcome pointed out that the applied LR model shows its effective classification outcome by obtaining maximum precision value of 98.60, recall of 98.60, F-measure of 98.60, ROC of 99.10, accuracy of 98.63 and kappa value of 96.29.