RiskVision

The Application

RiskVision is designed to assess patient risk for critical diseases based on structured questionnaires. The frontend, built in Next.js, securely collects user data and forwards it to a Flask backend.

The Models

The core predictive engine utilizes a custom-trained multi-layer neural network, supplemented by other classical ML models for baseline comparisons. Models were trained on publicly available datasets from Kaggle. A secondary Streamlit application was integrated to provide deep insights into model metrics, performance evaluation, and feature importance.