Abstract

Brain stroke is a major cause of death and long-term disability worldwide, placing disproportionate stress on healthcare systems and society. Why? A stroke diagnosis that is early can help prevent strokes, provide prompt treatment for those with high risk, and improve patient outcomes. Even so, traditional methods for assessing stroke risk often rely heavily on manual assessment, clinical experience, and limited rule-based scoring systems, which may not accurately account for intricate interactions among various risk factors. The paper describes an online platform that employs machine learning techniques to measure an individual's vulnerability to brain injury. Using supervised learning models, the proposed system can predict stroke risk levels by utilizing demographic data from individuals with different age groups and lifestyle factors. It uses a pipeline of structured data preprocessing and feature engineering to normalize inputs with missing values, improve predictive performance. Multiple machine learning models are trained and evaluated, with the top-performing model then being deployed within a web platform for real-time risk assessment. Compared to traditional statistical methods, the machine learning-based approach is proven to provide high accuracy and robustness in making predictions. By enabling scalable, data-driven stroke risk evaluation and decision-making, the platform aims to promote public health awareness and improve outcomes for both individuals and society.