Smart India Hackathon grand fianlist.
The goal of this project is to identify the similarity between given patients base on their demographics, lab-test values, prescription notes, discharge summaries etc.
I have used two different approached in order to predict the similarity-
- Use an autoencoder to get an enriched representation of the cleaned data and then pass it through another final autoencoder for
predictions - Use latent semantic indexing on the cleaned data.
The dataset is confidential as it contains seneitive medical data, Contact [email protected] regarding further queties.