Advanced Machine Learning Approaches for Heart Disease Prediction: A Study on Ensemble and Hybrid Models

The increasing availability of healthcare data has enabled machine learning (ML) models to play a crucial role in disease prediction and diagnosis. This research focuses on improving predictive accuracy by implementing hybrid models through stacking classifiers and comparing their performance against traditional supervised learning algorithms. The dataset, sourced from Kaggle, was preprocessed and subjected to multiple classification techniques, including Logistic Regression, Decision Trees, Random Forest, XGBoost, and Support Vector Machines. The hybrid model was constructed using a combination of base classifiers and a meta-classifier to enhance prediction robustness. Performance evaluation metrics such as accuracy were utilized to assess the effectiveness of the models. Experimental results demonstrate that the hybrid stacking classifier significantly outperforms individual algorithms, achieving higher accuracy and improved generalization capabilities. The findings of this study highlight the potential of ensemble learning techniques in advancing machine learning applications in healthcare.