Early Stroke Prediction by Machine Learning

Abstract

Stroke is a major global health concern and one of the leading causes of mortality and long-term disability. Early prediction of stroke risk is essential for timely medical intervention and prevention of severe complications. This paper presents a machine learning-based predictive framework designed to assess the likelihood of stroke occurrence using patient health records. The proposed system utilizes key clinical and lifestyle parameters such as age, hypertension, heart disease, glucose level, body mass index (BMI), and smoking status. Several classification algorithms, including Logistic Regression, Decision Tree, Random Forest, and Support Vector Machine (SVM), are implemented and evaluated for performance comparison. The proposed model is evaluated using a healthcare stroke dataset containing demographic and clinical attributes of patients. Experimental analysis shows that the Random Forest classifier achieved the highest prediction accuracy of 96.2% compared to other machine learning algorithms. The system incorporates data preprocessing techniques, feature engineering, and model evaluation using standard performance metrics such as accuracy, precision, recall, and F1-score. Experimental results indicate that ensemble methods, particularly Random Forest, outperform other models in terms of predictive accuracy and robustness. The proposed framework aims to support healthcare professionals in early diagnosis, risk assessment, and preventive care strategies.