AI health Diagnosis Assistant

Inspiration## Inspiration

Many people delay visiting a doctor simply because they don't know whether their symptoms are serious enough to warrant it. We wanted to build something that gives people an immediate, accessible first read on their symptoms enough to nudge them toward timely medical care instead of putting it off.

What it does

AI Health Diagnosis Assistant takes user-reported symptoms as input and uses a trained Decision Tree / Random Forest model to predict likely diseases and suggest basic precautions. It's built for early awareness, not diagnosis ,the goal is to help users understand what might be going on and encourage them to see a doctor sooner rather than later.

How we built it

• Data: Collected a symptom-disease dataset from Kaggle
• Preprocessing: Cleaned and prepared the data using Pandas and NumPy
• Modeling: Trained and tested a Decision Tree / Random Forest classifier with Scikit-learn
• Interface: Built a simple, non-technical front end with Streamlit so anyone can input symptoms and get results instantly
• Flow: Symptom input → ML model prediction → predicted disease + precautions → displayed to the user

Challenges we ran into

• Finding a symptom-disease dataset that was clean and complete enough to train on reliably
• Balancing model accuracy with the risk of overconfident predictions
• Making sure the tool clearly reads as preliminary guidance, not a diagnosis, so it doesn't get misused

What we learned

How to take a real, unglamorous everyday problem (people avoiding the doctor) and translate it into a lightweight, working ML pipeline ,and how important it is to communicate a model's limitations clearly to end users.

What's next

• Expanding the symptom-disease dataset for better coverage and accuracy
• Adding multi-language support for wider accessibility
• Exploring integration with real medical guidance systems for higher-stakes use

Built With

• decision-tree
• machine-learning
• numpy
• pandas
• python
• random-forest
• scikit-learn
• streamlit