IoT-Enabled Smart Healthcare System for Heart Disease Prediction Using Deep Learning and Dimensionality Reduction

Accurate predictive diagnostic systems have become essential because heart disease conditions are rising in frequency. Healthcare advancements led to broader patient data inclusion in medical databases that enhances heart disease diagnosis effectiveness. The current databases for ischemic heart disease struggle with four major problems involving both feature selection difficulties and insufficient sample sizes along with distribution imbalances and missing data points. The authors present a smart healthcare system powered by IoT with deep learning and dimensionality reduction methods to achieve accurate heart disease predictions. Through wearable IoT devices patients can give health data about their blood pressure and heart rate and oxygen saturation level and these readings are processed using deep learning algorithms based on Convolutional Neural Networks (CNNs). The predictive analysis utilizes principal component analysis alongside long short-term memory networks to accomplish efficient data simplification and performance improvement. The detection capabilities of the Attention Bidirectional Gated Recurrent Unit model (ABiGRU) improve with the help of a Parameter Optimization Approach (POA) to select its hyper parameters. The Kaggle dataset simulation revealed the system produced results of 96.80% accuracy on Dataset-I while attaining 94.80% accuracy on Dataset-II in addition to obtaining high precision values and recall scores and F1-scores. These predictive modeling outcomes prove that the system works effectively thus establishing itself as an accurate diagnostic tool for early heart disease prediction and improved patient healthcare treatment and demonstrating promising incorporation possibilities in IoT and deep learning based customized healthcare.