ENHANCING HEART ATTACK PREDICTION ACCURACY BY FINDING BEST SVM PARAMETER VALUE WITH FUZZY LOGIC

Cardiovascular disease (CVD) is still one of the highest causes of death in the world. One of them is heart disease. It is important to detect heart disease as early as possible so that management with counselling and medicines can begin as soon as possible. Delayed diagnosis will lead to increased morbidity and mortality in patients. However, detecting heart disease at an early stage is not easy, so getting accurate predictions will really help patients get treatment more quickly. Making accurate predictions is a challenge considering the large amount of data that must be processed. One prediction method that can be used is the use of the SVM machine learning algorithm. The use of the SVM algorithm in making predictions cannot be separated from finding a hyperplane that can separate the existing feature set maximally. This research will use Fuzzy Logic to find the best SVM parameter value, so that the resulting hyperplane can be maximized and can provide better prediction results. This sutdy use Heart Disease Classification Dataset from Kaggle, which consists of 1.319 data that has eight input fields and one output field. Results of prediction heart attack risk from standard SVM shows 80% of whole accuracy with 73% precission for negative class and 83% precission for positive class. The whole process takes 1.362 seconds to finish. Results of prediction by using SVM with best prameter value from fuzzy logic give 89.39% whole accuracy, 85% precission for negative class and 92% precission for positive class. The whole process takes 1 minute 10 seconds to finish using all the computer CPUs. This research shows that using fuzzy logic to find the best SVM parameter value can improve the performance of the SVM algorithm in predicting heart attacks.

Keywords— enhancing prediction, cardio vascular disease, fuzzy logic, SVM