The random forest model achieved higher predictive performance for short-term heart attack risk compared to baseline models in analyzing IoT-connected cardiovascular health data.
An R-based machine learning pipeline effectively predicts short-term heart attack risk from IoT cardiovascular data, with random forests providing the highest accuracy and decision trees offering transparent clinical rules.
Absolute Event Rate: 0% vs 0%
Wearable sensors and IoT platforms enable continuous health monitoring, but the large volume of raw data they generate often overwhelms clinicians and remains underutilized. This paper presents an end-to-end R-based analytical workflow for analyzing IoT-connected cardiovascular health data, including resting blood pressure, cholesterol, heart rate, chest pain indicators, and age. The proposed system performs data preprocessing, normalization, and exploratory analysis, followed by the application of statistical learning models such as logistic regression, decision trees, and random forests to predict short-term heart attack risk. Clinician-friendly visualizations including histograms, scatter plots, ROC curves, and variable-importance plots are generated to improve interpretability. Experimental results show that the random forest model achieves higher predictive performance compared to baseline models, while decision trees provide transparent rule-based insights aligned with clinical reasoning. This work demonstrates the effectiveness of R for IoT healthcare analytics and provides a reproducible framework for clinical and home-care deployment.
Kumar et al. (Sat,) reported a other. The random forest model achieved higher predictive performance for short-term heart attack risk compared to baseline models in analyzing IoT-connected cardiovascular health data.