Web-Based Stroke Prediction System Using Machine Learning

Tan, Lock Kwan (2025) Web-Based Stroke Prediction System Using Machine Learning. Final Year Project (Bachelor), Tunku Abdul Rahman University of Technology and Management.

Text RDS_Tan Lock Kwan_Fulltext.pdf Restricted to Registered users only Download (5MB)

Abstract

Stroke is the leading cause of death and disability worldwide, and early detection is critical to mitigating its effects. This project creates a web-based stroke prediction system using machine learning to identify individuals at risk based on demographic, medical, and lifestyle information. This system provides users with actionable insights for early intervention and guides healthcare professionals through decision-making. The project aims to improve health outcomes and reduce the socioeconomic burden of stroke by providing a reliable and easily accessible stroke prediction tool. Preprocessing is performed on the dataset to handle missing values, address class imbalance, and encode categorical and continuous variables. Multiple machine learning models, such as Gradient Boosting, Random Forest, and Support Vector Machine (SVM), are evaluated using metrics such as accuracy and F1-score, with hyperparameter tuning applied to enhance model performance. Although Gradient Boosting achieved the highest overall accuracy (86.01%), the SVM model was selected for deployment due to its superior recall of 68% in detecting stroke cases, making it more effective at minimizing false negatives. The final model is deployed via an easy-to-use web platform, which allows users to input health-related data and receive stroke risk predictions. The system demonstrates strong performance in early risk detection and provides a user-friendly interface for both anonymous and registered users. However, its relatively lower precision for positive cases (18%) highlights a need for future improvements, such as increasing dataset diversity and enhancing model sensitivity, to further strengthen the reliability of predictions.