Machine Learning-Based Pre-Stroke Detection System

Stroke remains a significant health concern worldwide, often leading to severe disabilities and mortality. Early detection and intervention are crucial for preventing strokes and mitigating their impacts. In this study, the proposed machine learning-based pre-stroke detection system aimed at identifying individuals at risk of experiencing a stroke before the onset of symptoms. The proposed system utilizes a diverse dataset containing demographic information, medical history, lifestyle factors, and biomarkers relevant to stroke risk. Various machine learning algorithms, including Decision Tree Algorithm, Random Forest Algorithm, Naïve Bayes Algorithm, Multilayer Perceptron Algorithm, and J Rip Algorithm, were trained, and their evaluation was conducted separately using the dataset. The proposed system uses a cloud-based stroke warning system that utilizes machine learning technology to predict impending strokes. The peculiar scrutiny of many machine learning methods led to the development of an effective machine learning strategy for the precise detection of strokes. A microcontroller-based system was used to establish and exhibit an actual patient monitoring system that can sense a variety of real-time indicators, including core temperature, blood pressure, blood flow, heart rhythm, and oxygen level, to let caregivers and doctors keep an eye on stroke patients around the clock. Decisions can be made quickly and easily with the aid of various decision-making algorithms, and anyone can access the database according to their needs.