Heart Disease Prediction using Optimized Feature Set and Classifiers

Heart disease has emerged as one of the most dangerous illnesses, significantly affecting people's quality of life. To save patients from suffering an accurate and on-time diagnosis is extremely crucial. With the help of clinical parameters and machine learning this process can be made less complex. This paper compares different types of meta-heuristic feature selection techniques, ten types of machine learning classifiers and Deep Belief Network classifier to compare the performance metrics. This process is carried out for two different datasets namely Cleveland and Statlog dataset obtained from UCI machine learning repository and a maximum accuracy of 88.87% is achieved for Cleveland dataset and 88.88% for Statlog dataset.