pqy000 / DiagNCF

DiagNCF

A pytorch GPU implementation of #016095 al. "Diagnosis Neural Collaborative Filtering for Accurate Medical Recommendation" of Qualification.

The requirements are as follows:

* python==3.6
* pandas==0.24.2
* numpy==1.16.2
* pytorch==1.0.1
* gensim==3.7.1
* tensorboardX==1.6 (mainly useful when you want to visulize the loss, see https://github.com/lanpa/tensorboard-pytorch)

Example to run:

python main.py

Dataset

We provide two processed datasets: MIMIC3-20 and MIMIC3-30

train.rating:

• Train file.
• Each Line is a training instance: ICD9_new\t ITEM_new\t COUNT_new (if have)

test.rating:

• Test file (positive instances).
• Each Line is a testing instance: ICD9_new\t ITEM_new\t COUNT_new (if have)

test.negative

• Test file (negative instances).
• Each line corresponds to the line of test.rating, containing 99 negative samples.
• Each line is in the format: (ICD9_new,ITEM_new)\t negativeItemID1\t negativeItemID2 ...

Pretrained weights

The trained model weights are under "models/"