AI/ML-Based IoT Health Monitoring and Risk Prediction Using an ESP32 Sensor Node, Random Forest Classification, and a Streamlit Dashboard
Department of Artificial Intelligence and Data Science, Ajeenkya DY Patil School of Engineering, Pune, India · India
Abstract
Health problems are easier to manage when abnormal physiological trends are noticed early.. Right now routine monitoring still depends mostly on periodic manual checks. These checks happen at set times rely on the person doing them and are hard to use in homes or remote areas. This paper introduces Smart Health Monitor, a low-cost prototype that brings sensing, machine learning and visualization together in one process. An ESP32 microcontroller collects data from sensors and sends it through a serial link to a Python application. That application cleans the data loads a Random Forest classifier that was trained before and gives a health-risk class: low, medium or high. A Streamlit dashboard shows the real-time readings along with the predicted risk level. The paper looks at work on IoT-based healthcare, wearable monitoring and machine learning models that predict health risks. It points out the gap between showing raw values and offering affordable real-time analysis. It also outlines a method for training models comparing them and validating results. The prototype is meant as a decision-support tool and a demonstration system. It is not a certified device. No performance numbers are given here. All numerical results will be shared after the experiments, in Section VIII are completed.
Keywords
How to Cite This Article
Rights
BODHIVRUKSHA JOURNAL OF DIVERSE DISCIPLINE