Heart Disease Prediction Using Ensemble Feature Selection Method and Machine Learning Classification Algorithms

One of the most serious diseases in the present human society is cardiovascular disease. This illness strikes a person very suddenly, leaving people with little opportunity to receive treatment. Therefore, it is quite challenging for clinical diagnostics to accurately identify patients at the appropriate time. Using an efficient heart disease prediction model, cardiovascular disease can be identified, and the treatment can be provided quickly to save human life. In this study, using the novel frequent features subset selection, the features which are most relevant are selected. The classification methods like decision tree, K‐nearest neighbor, random forest, and gradient boosting are applied to the dataset with the selected features. The proposed model accuracy is compared with the accuracy of the model using backward selection and the model using recursive feature elimination. Finally, it was proven that the proposed model worked effectively and had better accuracy than the other models.