CardioPredictor: An Intelligent Risk IoT-Based Classification Model for Cardiovascular Disease in Low-Resourced Environments

Cardiovascular disease (CVD) is the leading global cause of death, with low- and middle-income countries most affected due to limited diagnostic infrastructure. Although machine learning models have shown strong predictive performance, few have been deployed on wearable, low-power systems suitable for resource-limited environments. This study presents CardioPredictor, an intelligent system for real-time CVD risk classification using electrocardiogram (ECG) data and machine learning, deployed on an Internet of Things (IoT) device. The dataset comprised 4,997 heartbeat records (2,079 normal and 2,918 abnormal). Seven supervised models including, XGBoost, Support Vector Machine (SVM), K-Nearest Neighbour (KNN), Decision Tree, Naïve Bayes, Multilayer Perceptron (MLP), and Linear Discriminant Analysis (LDA) were trained and evaluated using stratified 10-fold cross-validation with accuracy, precision, recall, and F1-score metrics. XGBoost achieved the best results, with 99.9% accuracy, 99.8% precision, 99.5% recall, and 99.4% F1-score and was therefore deployed onto an Arduino UNO R4 WiFi platform using MATLAB Coder. Real-time inference produced results within 0.3 seconds, displayed on an OLED screen and consistent with clinical instruments.