Brain Stroke Prediction Using Machine Learning

A stroke, or cerebrovascular accident (CVA), is a critical medical event resulting from disrupted blood flow to the brain, often causing permanent damage. Understanding its causes, types, symptoms, risks, and prevention is crucial, as it stands as the leading cause of death and disability. There are three main types: Ischemic stroke (87% of cases, caused by artery blockage), Hemorrhagic stroke (from a ruptured blood vessel), and Transient Ischemic Attack (TIA), a temporary block-age. Ischemic strokes are primarily caused by atherosclerosis or embolism. Hemorrhagic strokes can result from conditions like aneurysms or high blood pressure. Symptoms include sudden weakness, numbness, difficulty speaking, vision problems, headache, and lack of coordination. Risk factors encompass high blood pressure, smoking, diabetes, obesity, family history, age, and heart disease. Preventive measures involve managing blood pressure, cholesterol, and diabetes, quitting smoking, maintaining a healthy lifestyle, and limiting alcohol consumption. Stroke survivors may face physical, cognitive, or emotional challenges, but treatment and ongoing support can enhance their quality of life. Ongoing research aims to improve prevention, treatment, and recovery, with telemedicine and medical technology accelerating these processes. This article proposes the use of machine learning algorithms (decision tree, naive Bayes, K-nearest neighbor, Random forest, logistic regression) to create a prediction model for brain strokes. The decision tree algorithm is deemed the best, achieving an overall accuracy of 90.4%. The subsequent step involves building a website using Flask for model deployment. Early diagnosis significantly impacts recovery, and the integration of technology in healthcare, such as machine learning, holds promise in stroke prediction and management.