Advanced Hybrid Learning Architecture for Precision Cardiovascular Risk Assessment

Abstract

Remaining the major motive of demise worldwide, cardiovascular diseases (CVDs) spotlight the urgent need of accurate computational threat prediction fashions. The use of a complete heart disease Dataset obtained from 1,190 patients across 5 different databases, this work presents a twin-stage stacked machine learning (ML) approach for predicting heart diseases. Eleven important traits within the dataset are absolutely essential for risk evaluation. Randomized seek CV and Grid seek CV were used for hyperparameter optimization among all algorithms in order to guarantee strong model performance. To improve forecast accuracy, the proposed framework combines several ML classifiers consisting of ensemble techniques. With an accuracy of 99.0%, a voting Classifier including a Bagging Classifier with Random forest (RF) and decision Tree (DT) confirmed better overall performance. This result emphasizes how well combined approaches handle the complexity of coronary heart disease prediction. Early intervention and higher clinical decision-making in cardiovascular healthcare are made possible by the suggested scalable and dependable answer.