Enhancing Heart Disease Prediction with Ensemble Deep Learning and Feature Fusion in a Smart Healthcare Monitoring System

In order to improve the early diagnosis of heart illnesses, this study investigates the use of cutting-edge machine learning algorithms within the context of a smart healthcare monitoring system. To continuously collect and upload patient data to a healthcare database, various physiological sensors—such as temperature, pressure, blood glucose, oxygen saturation, and ECG sensors—are employed. Machine learning models, including convolutional neural networks (CNN), recurrent neural networks (RNN), and random forests (RF), process the comprehensive dataset encompassing relevant indicators and trends. Real-time trend and outlier data analysis using these models are crucial for swift medical assistance and early detection of heart issues. The study concludes with an ensemble model approach that combines the adaptability of conventional algorithms (RF) with the advantages of deep learning models (CNN and RNN). Demonstrating impressive accuracy rates of 98% for CNN, 94.5% for RNN, and 91.11% for RF, the findings highlight the models’ precision. Evaluation metrics for recall, precision, and F1 score provide a comprehensive assessment of the models’ performance, particularly in reducing false negatives—an essential consideration in the healthcare sector. In the era of advanced monitoring, this study underscores the revolutionary potential of data-driven healthcare. Beyond emphasizing the importance of early disease prediction, it paves the way for integrating machine learning into healthcare to enhance patient care, interventions, and service delivery. This research constitutes a crucial step toward proactive, personalized treatment and improved patient outcomes in the rapidly evolving healthcare landscape.