Stroke Prediction System using Machine Learning

• Shubham Banwal

• April 2025

Stroke Risk Prediction System using Machine Learning & Big Data

Problem:

Stroke is a leading cause of mortality and disability, yet early risk prediction remains challenging due to complex, multi-factor health data and lack of scalable predictive systems.

Approach:

Developed a machine learning–based predictive system that analyses large-scale patient data to identify individuals at high risk of stroke, enabling early intervention and preventive decision-making.

System Design:

• Data ingestion and preprocessing pipeline for patient health records
• Feature engineering layer (e.g., age, medical history, lifestyle indicators)
• Machine learning models for risk classification
• Big data processing framework for handling large datasets
• Evaluation and validation framework for model performance

Key Contributions:

• Built data pipelines to clean, process, and prepare large-scale healthcare datasets
• Engineered features to capture key risk indicators for stroke prediction
• Implemented machine learning models (classification-based) to predict stroke risk
• Evaluated model performance using metrics such as accuracy, precision, and recall
• Addressed data imbalance and improved model robustness
• Generated interpretable outputs to support healthcare decision-making

Constraints & Tradeoffs:

• Data quality and missing values in healthcare datasets
• Class imbalance (fewer stroke cases vs non-stroke cases)
• Tradeoff between model accuracy and interpretability
• Ethical considerations in handling sensitive health data