Morality Prediction Model in Cardiovascular Disease with Significant Feature Selection and Hybrid KNN Classification Technique

Abstract

Nowadays morality rate is increasing globally due to heart disease. It is one of the leading health risk facing men today. So early detection of heart disease assist the patients to maintain a healthy life style. Several techniques are used in the medical field to detect or diagnose disease in view of patient family health history and some other aspects. However, developing a system to predict the heart diseases without any medical tests is still challenging. Machine learning (ML) approaches is suitable and effective in providing decision and prediction from enormous health care data. Several previous researches provide an overall view in ML methods for disease prediction but the accuracy of prediction is still needed to be improved. In this study, a novel framework is presented that intent at removing the unwanted features with Bacterial Colony Optimization algorithm and applies the Hybrid KNN algorithm with great accuracy in identifying the heart disease. This prediction model is developed with UCI Cleveland dataset with several known classification approaches. An enhanced model is presented with 99.83% accuracy in heart disease prediction. The presented study is compared with other classification approaches.