This repo includes the codes for the submission: Breaking Student-Concept Sparsity Barrier for Cognitive Diagnosis

Our environment:

Python 3.9.7 
torch 2.0.1
pandas 1.3.4
scikit-learn 0.24.2
networkx 2.6.3

1. Run codes on ASSIST dataset

cd assist/codes
python main_our.py

We also include two representative baseline models. Please refer to main_NCDM.py and main_KaNCD.py for more details.

2. Run codes on MOOC-Radar dataset

This dataset is too large to upload in a repo. Please refer to https://github.com/THU-KEG/MOOC-Radar and download the coarse version.

Please first run "python divide_data.py" to preprocess data.

Then, you can run the codes by

cd mooc/codes
python main_our.py