imposter404/Co-Disease-Prediction ? reverse-engineered prompt

Build me a simple Python project that can predict possible co existing diseases from a patient’s symptoms and past disease history. I want it to use the data already in this repo and combine the current approaches so it can look at symptoms, check common disease pairs from historical data, and give a final list of likely diseases instead of just one.

Please make it easy to run and understand. A person should be able to test all patients in the dataset or enter one patient and see the predicted diseases, the percentage or confidence for each one, and the model accuracy. Keep the current ensemble style idea with the different models and the association rule approach, but clean up the code so it actually works end to end.

If possible, add a very basic web page in the Django part where someone can enter symptoms and view the results, but the script and notebook should still run too. Also fix the requirements, data loading, and any broken naming or output issues. Look up current docs online if you need to.