DearLLM: Enhancing Personalized Healthcare via Large Language Models-Deduced Feature Correlations

Welcome to the official repository for DearLLM: Enhancing Personalized Healthcare via Large Language Models-Deduced Feature Correlations.

📢 News: this work has been accepted at the AAAI 2025 !

If you find our project interesting or helpful, we would appreciate it if you could give us a star! Your support is a tremendous encouragement to us!

Requirements

All dependencies are described in the file requirements.txt, you can install the packages required for this project via the command below.

pip install -r requirements.txt

Usage

Prepare Dataset

As we can not provide the MIMIC-III and MIMIC-IV datasets, you must acquire the data yourself from https://mimic.physionet.org/. Please reference https://github.com/sunlabuiuc/PyHealth for more details about data preparation and put the processed data into data folder before entering into the folllowing steps.

1. Generate needed information from data_preprocess folder

   python data_preprocess/utils.py

2. Generate relation graph from the diagnoses result of each patient by code under perplexity-server folder and perplexity-client folder.

   python perplexity_server/server.py
   python perplexity_client/client.py

3. Postprocess perplexity graphs followed by instructions in data_preprocess/data_postprocess.ipynb, which normalize perplexity score into range [0, 1].

4. Merge patient sample data, node feature and graph information into one file.

   python data_preprocess/dataset_merge.py

Run Code

To run the code, just run main_dearllm.py in root folder.

python main_dearllm.py