Heart Disease Prediction: A Comparative Analysis of Machine Learning Algorithms

Nowadays, heart disease is one of the biggest concerns. A WHO statistic states that 17.9 million people worldwide die each year, accounting for 32% of all deaths worldwide. It is now very difficult to diagnose and start therapy at an early stage due to population growth. In the healthcare industry, earlier machine learning techniques have been highly successful. The study focuses on predicting cardiac disease using historical data and knowledge. Much greater precision, correctness, and perfection are needed in the analysis and prognosis of cardiac-related issues because, if left undiagnosed, the condition can be fatal. We require a crucial prediction system to address such an issue. This study calculates and determines how accurately machine learning algorithms predict cardiac disease. A variety of machine learning algorithms, including Random Forest Classifier (RF), Neural Network (MLP), Logistic Regression (LR), Naïve Bayes (NB), Support Vector Machine (SVM), Decision Tree (DT), and K-Nearest Neighbour Classifier (KNN), are used to make this prediction. Training and testing of these algorithms are done using the heart dataset. In training, 80% of the dataset is used, and 20% of the dataset is used for testing. The metrics Accuracy, F1-Score, Recall, Precision, and ROC-curve are used for comparison. The results show that RF, MLP, MLP,RF, and RF have the highest accuracy (86.96), F1-score (86.79), recall (0.91), precision (0.82), and ROC-curve (0.93), respectively.