Stroke Prediction Using Machine Learning

Abstract

Stroke, a cerebrovascular event, represents a significant global health concern due to its substantial impact on morbidity and mortality. It occurs when there is a sudden interruption or reduction of blood supply to the brain, leading to the impairment of brain function. As the second leading cause of death globally, stroke demands urgent attention, and early detection is pivotal for effective intervention. This study addresses the global health concern of strokes by leveraging machine learning models for early detection and risk assessment. The study employs logistic regression, random forest, naive Bayes, and support vector machine algorithms to create a robust predictive model. Key objectives include data cleaning, addressing class imbalance, model evaluation, and deployment. The research contributes to the growing literature on machine learning applications in healthcare by presenting a holistic approach to stroke prediction. Results indicate that while random forest achieves high accuracy, logistic regression provides a balanced sensitivity-specificity trade-off. The models are deployed through an interactive Shiny app, enhancing accessibility and usability for healthcare professionals. Future work involves refining models, incorporating additional features.