This repository is for our submitted paper for MDPI-Biology '[Integration of Multimodal Data from Disparate Sources for Identifying Disease Subtypes]'.
This repository is based on PyTorch 1.4 and CUDA 10.0.
For installing PyTorch combining with the CUDA toolkit, please follow the official instructions in here. The code is tested under PyTorch 1.4 and Python 3.6 on Ubuntu 18.04.
- Download this repository. And go to folder:
cd code
-
Download the data here and put in the data folder. An illustration of the folder organization is:
-root --data ---gbm ---laml ---paad
-
Train and test the model.
Run complete fusion model (CFA in the paper):python main_aec.py
Run incomplete fusion model (IFA in the paper):
python main_aec.py --pred_missing
Run complete fusion model with 2 existing modalities (CFA-2M in the paper):
python main_aec_2m.py
Run single modality model (SMA in the paper), change which modality to use
m_use = m*
inmain_aec_single.py
:python main_aec_single.py