ChronicRisk AI

Inspiration

Chronic diseases remain a major global health burden. Early risk factors exist but are often underutilised. Healthcare systems focus more on treatment than prevention. We wanted to shift the focus toward proactive, data-driven screening.

What it does

ChronicRisk AI predicts early risk of chronic diseases. It uses routine clinical parameters like BMI and glucose. The system classifies patients into low, medium, or high risk. It supports clinicians in preventive decision-making.

How we built it

We used a public, de-identified clinical dataset. Data was cleaned, normalised, and preprocessed. We trained logistic regression and random forest models. Performance was evaluated using accuracy and sensitivity.

Challenges we ran into

Handling missing and inconsistent clinical data. Balancing interpretability with model performance. Preventing overfitting with limited features. Addressing potential demographic bias.

Accomplishments that we're proud of

Built a working ML-based risk prediction prototype. Designed a simple and interpretable classification system. Aligned the project with preventive healthcare goals. Maintained strong ethical considerations throughout development.

What we learned

Data preprocessing is critical in healthcare ML. Simple models can perform surprisingly well. Interpretability is essential in clinical tools. Ethical responsibility is central to medical AI.

What's next for ChronicRisk AI

Validate the model on larger datasets. Incorporate explainable AI techniques. Expand predictions to multiple chronic diseases. Develop a user-friendly clinical dashboard.

Built With

• matplotlib
• numpy
• pandas
• python
• scikit-learn