A Comparison of Supervised Learning Algorithms to Prediction Heart Disease

Abstract

Heart disease problems are rapidly increasing day to day. Humanslose their lives at an early stage. Consequently, themain purpose of this projectis to employ supervised machine learning methods for heart disease early prediction. For the prediction and diagnosis of cardiac diseases, different techniques are used like data mining and machine learning. This would be tremendously useful to human life since, owing to a lack of cardiovascular competency and quick development in improperly diagnosed instances, heart diseases in people might develop at an early stage. As a result, developing robust and effective early-stage cardiac illness prediction by using analytical decision-making and digital patient data might alleviate this problem. To predict heart diseases, numerous supervised machine-learning techniques were used to learn about the illness, and their efficiency and accuracy were evaluated. This study used a Kaggle dataset on heart disease and found that three classification methods-, K-Nearest Neighbor (KNN), Support Vector Classifier, and Multi-Layer Perception (neural network) could accurately classify heart disease. TheKNN is given 91.8% accuracy. As a result, we discovered that KNN results can more accurately forecast the chance of patients developing heart disease.