COMPARATIVE ANALYSIS OF DIFFERENT DATA MINING TECHNIQUES FOR PREDICTING THE RISK OF HEART DISEASE

ABSTRACT

Heart disease is one of the significant causes of human death worldwide. Therefore, accurate prediction of this disease in individuals is paramount to providing early intervention and preventing the risks associated with complications and health issues. Data mining techniques have shown great promise in predicting the likelihood of heart disease based on data analysed from patients' records. In this study, four different classifiers - random forest, multilayer perceptron, naïve bayes, and J48 were experimented with and implemented in the WEKA tool to compare the best model using standard performance metrics. The results of the comparative analysis showed that random forest outperformed the other three classifiers with an accuracy of 100.0%. The findings indicated that a random forest classifier could effectively predict whether an individual will either or not have heart disease, given sufficient case data.

Keywords: Comparative analysis, Data Mining, Heart disease, Prediction, WEKA