<![CDATA[Heart disease is one of the leading causes of death in the world, including in Indonesia, which ranks second after stroke. Early detection is essential to reduce the risk of serious complications and death from cardiovascular disorders. This study aims to design an Internet of Things (IoT)-based early detection system for heart health that is integrated with MAX30102 sensors and Random Forest algorithms to classify heart rate conditions. Biometric data in the form of heart rate (BPM), blood oxygen level (SpO₂), and activity condition features (rest, light exercise, stress) were collected from 150 respondents. This data collection was validated by comparing the results using ECG devices by medical personnel. Pre-processing is done through data cleansing, category variable encoding, and feature extraction (BPM variability, PPG amplitude). The classification model was developed with the Random Forest 100 decision tree and tested with 5-fold cross validation. The results showed that the system was able to achieve an average accuracy of 93% with a standard deviation of 0.03, as well as an accuracy per fold of 93%, 93%, 97%, 93%, and 87%. The classification results are in line with the ECG data of medical personnel, indicating that this system is reliable enough for the early detection of normal or abnormal heart conditions. The study concluded that the integration of IoT and Random Forest is effective as a real-time, cost-effective, and supporting early detection of heart health, especially in remote areas. Advanced development is suggested to expand activity data and add biometric features to improve classification accuracy.]]>
Copyrights © 2025
