Temporal Enhanced Inductive Graph Knowledge Tracing

IGKT-GAT Applied Intelligence 2023

@article{han2023temporal,
  title={Temporal enhanced inductive graph knowledge tracing},
  author={Han, Donghee and Kim, Daehee and Kim, Minsu and Han, Keejun and Yi, Mun Yong},
  journal={Applied Intelligence},
  pages={1--18},
  year={2023},
  publisher={Springer}
}

Datasets

• EdNet dataset from : https://github.com/riiid/ednet
• ASSIST2017 dataset from : https://sites.google.com/view/assistmentsdatamining

Download datasets

1. make up : build docker image and start docker container
2. python3 download_datasets.py : download datasets

Docker Container

• Docker container use igkt project directory as volume
• File change will be apply directly to file in docker container

Preprocessing

1. make up : build docker image and start docker container
2. python3 src/pre_process.py : start ednet data preprocessing in docker container
3. python3 src/pre_process_assist.py : start assist data preprocessing in docker container
4. python3 src/item_preprocessing.py : start ednet item data preprocessing in docker container
5. python3 src/item_preprocessing_assist.py : start assist item data preprocessing in docker container

Train

1. make up : build docker image and start docker container
2. check train_config/train_list.ymal file (default: assist2017 with igkt_ts and igkt_gat)
3. python3 src/train.py : start train in docker container