🌟 Inspiration
Healthcare delays and lack of quick medical guidance often lead to late diagnosis. Many people search symptoms online but get unreliable or overwhelming results. The goal was to build an accessible, AI-powered tool that provides instant, data-driven health insights—not as a replacement for doctors, but as a first-step helper for medical awareness.
🔍 What it does
The Smart Health Diagnosis Web App allows users to enter their symptoms and instantly receive predicted possible diseases using a trained Machine Learning model. It provides: ✅ Symptom-based disease prediction ✅ Simple and clean web interface ✅ Fast response powered by a trained Random Forest model ✅ Cloud-ready deployment for global access
🛠️ How we built it
Frontend: Built using HTML, CSS, and Flask templates for a clean user experience Backend: Flask API handling inputs and model predictions Model: Random Forest Classifier trained using scikit-learn on a labeled symptoms–disease dataset Data Processing: Pandas + NumPy Model Storage: Saved with joblib Folder structure included: model training scripts, saved .pkl files, and web app integration.
🚧 Challenges we ran into
Mapping diverse symptoms into meaningful ML features Ensuring the model remained accurate and didn’t overfit Handling environment differences between local and cloud deployment Keeping UI simple while processing multiple symptoms dynamically Dealing with missing or inconsistent dataset values
🏆 Accomplishments that we're proud of
Successfully integrated a trained ML model into a live web app End-to-end working pipeline from dataset → training → deployment Built a usable, real-world health tool instead of just a notebook model Achieved stable prediction results using Random Forest Designed a structure ready for scaling with APIs and future features
📚 What we learned
Practical ML deployment beyond just training models Importance of proper preprocessing and model saving Flask + ML integration workflows Handling user input variations for real-world data applications
🚀 What's next for AI-Powered Smart Health Diagnosis Web App
🔹 NLP-based natural language symptom input (no checkbox lists) 🔹 Larger dataset with more diseases & severity levels 🔹 Geolocation-based nearby hospital or doctor suggestion 🔹 Mobile app version (Flutter / React Native) 🔹 Multi-language support for accessibility 🔹 Early-stage health recommendation system (diet, remedies, precautions) 🔹 API version so other apps can integrate the model
Log in or sign up for Devpost to join the conversation.