Stroke Risk Prediction — Machine Learning & Interactive Dashboard

Project Overview

Stroke Risk Prediction is a machine learning project that predicts potential stroke risk based on demographic and clinical characteristics.

The project combines a Naive Bayes classification model with an interactive Streamlit dashboard for real-time predictions and data exploration.

Data Processing

Performed data preprocessing and preparation, including:

• Feature encoding
• Target selection
• Train-test splitting
• Preparation of numerical and categorical features

Machine Learning Model

Developed and evaluated a Naive Bayes classification model to predict potential stroke risk.

Hyperparameter tuning was performed to improve model performance, followed by comparison of model performance before and after tuning using classification evaluation metrics.

Interactive Dashboard

Developed an interactive Streamlit dashboard that allows users to:

• Input new patient data
• Generate real-time stroke risk predictions
• Explore dataset visualizations
• Review model performance