Hybrid Machine Learning and Deep Learning Approach for Heart Attack Prediction Using Clinical, Lifestyle, and Time-Series Data with Enhanced Feature Selection and Classification

Abstract

Cardiovascular diseases (CVDs), especially heart attacks, remain a leading cause of death globally. Early prediction and intervention are crucial to improving survival rates. This research proposes a hybrid machine learning and deep learning framework for heart attack prediction, utilizing real-time data collected through Internet of Things (IoT) sensors. The framework integrates Random Forest and Long Short-Term Memory (LSTM) models to analyze both structured health data (age, blood pressure, cholesterol) and time-series ECG signals. The hybrid approach enhances predictive accuracy by combining the strengths of ensemble learning and deep learning. Furthermore, privacy concerns are addressed through federated learning and secure data transmission. The model outperforms traditional methods, achieving high accuracy, recall, precision, and AUC, demonstrating its potential for real-time heart attack detection. This system offers a scalable, secure, and interpretable solution for cardiovascular disease prediction in diverse healthcare settings, ultimately contributing to better patient outcomes.