ARCHIVES

Abstract: This model aims to predict stroke risk using a dataset containing features like age, BMI, glucose levels, and lifestyle factors. The goal is to build a predictive tool that identifies individuals at high risk of stroke using machine learning models. The primary challenge of class imbalance is addressed using SMOTE, which enhances model performance on the minority class.Both Logistic Regression and Random Forest models were trained, with Random Forest outperforming Logistic Regression. Random Forest algorithm achieves high accuracy, precision and an AUC score, when comparing with Logistic Regression. Overall, Random Forest is a more accurate and reliable tool for stroke prediction.