The code is related to the paper
GIKT: A Graph-based Interaction Model for Knowledge Tracing published at ECML-PKDD 2020 Paper in arXiv.
You can download the preprocessed dataset: download
If you want to process the datasets by yourself, you can download them by the following links:
You should create a folder named
data for datasets. Besides, folders
logs are also needed to create for saving model and log in our experiments.
python == 3.6.5
tensorflow == 1.12.0
numpy == 1.15.2
Examples to run the model
python main.py --dataset assist09_3 --n_hop 3 --skill_neighbor_num 4 --question_neighbor_num 4 --hist_neighbor_num 3 --next_neighbor_num 4 --model hsei --lr 0.001 --att_bound 0.7 --sim_emb question_emb --dropout_keep_probs [0.8,0.8,1]
If you have more questions about our experiments, you can contact us. email: firstname.lastname@example.org