Predictive Modelling for Ischemic Brain Stroke Prediction Using Ensemble of Machine Learning Techniques

Abstract

According to the World Health Organisation [WHO], more than 15 million people globally suffer a stroke each year, which leads to approx. 5 million fatalities and severe disabilities of 30% of survivors. The CDC rank stroke as the fifth most common cause of death in the world. It’s important to prevent strokes in the first place because they usually happen during the first phase. Technologies that can detect brain strokes before symptoms appear are desperately needed to diagnose the condition early and prevent catastrophic consequences. To guarantee the accuracy and applicability of the model in the medical field, this project will prioritize ethical considerations, the interpretability of results, and collaboration with medical experts. This paper explores the application of machine learning techniques with an ensemble approach, such as Neural Networks, KNN, Naive Bayes classifier, Random Forest, and Logistic regression to predict ischemic brain strokes based on various risk factors like age, hypertension, diabetes, smoking, and cholesterol levels. The proposed models aim to achieve high accuracy, sensitivity, and specificity in stroke prediction, offering healthcare providers a robust decision-making tool. A comparative analysis of these machine learning models is conducted to determine the most effective approach in predicting ischemic brain stroke, focusing on reducing false positives and improving the precision of early diagnosis. The results demonstrate the potential of machine learning in transforming stroke prediction, contributing to better patient outcomes and more efficient use of healthcare resources.