Efficient Human Heart Disease Detection using Machine Learning Algorithm

Abstract

This project pioneers a novel approach to predicting cardiac disease using Machine Learning algorithms like LR, KNN, SVM, GBC, and the powerful Extreme Gradient Boosting Classifier (XGBoost) with GridSearchCV. Utilizing 5-fold cross-validation, it assesses performance across diverse datasets. The XGBoost Classifier with GridSearchCV achieves outstanding accuracy, hitting 100% in testing and 99.03% in training across multiple datasets, outperforming other algorithms and previous studies. Notably, the XGBoost Classifier without GridSearchCV also demonstrates strong accuracy. Furthermore, an extension employing Random Forest shows comparable accuracy with reduced computation time. This research underscores the efficacy of the proposed technique in advancing cardiac disease prediction.