COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS FOR STROKE PREDICTION

Abstract

In the modern world, data mining is crucial to the medical industries' ability to forecast disease. A stroke can harm many different body organs. Over 110 million people worldwide suffer from strokes. In India, stroke ranks as the fourth most prevalent cause of death. Every 4 minutes, a stroke victim passes away, although up to 80% of strokes may be avoided if we could detect or anticipate a stroke in its early stages. According to the National Stroke Association, hemorrhagic strokes cause around 40% of all stroke deaths. The application of machine learning algorithms in early disease diagnosis is revolutionizing the field of health care. To determine if the patient had a stroke or not, supervised algorithms including Logistic Regression, K-Nearest Neighbour, Naive Bayes (NB), Support Vector Machines (SVM), Decision Tree, Random Forest, and XG Boost were utilized. Out of all the algorithms used, the highest accuracy of 95.98 % was obtained using XG Boost.