LPKT_tensorflow_version

Source code and data set for the paper Learning Process-consistent Knowledge Tracing.

The code is the implementation of LPKT model, and the data set is the public data set ASSIST2012-2013.

If this code helps with your studies, please kindly cite the following publication:

@inproceedings{10.1145/3447548.3467237,
author = {Shen, Shuanghong and Liu, Qi and Chen, Enhong and Huang, Zhenya and Huang, Wei and Yin, Yu and Su, Yu and Wang, Shijin},
title = {Learning Process-Consistent Knowledge Tracing},
year = {2021},
isbn = {9781450383325},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3447548.3467237},
doi = {10.1145/3447548.3467237},
pages = {1452–1460},
numpages = {9},
location = {Virtual Event, Singapore},
series = {KDD '21}
}

Dependencies:

• python >= 3.7
• tesorflow-gpu >= 2.0
• numpy
• tqdm
• utils
• pandas
• sklearn

Usage

First, download the data file: 2012-2013-data-with-predictions-4-final.csv, then put it in the folder 'data/'

Then, run data_pre.py to preprocess the data set, and run data_save.py {sequence length} to divide the original data set into train set, validation set and test set.

python data_pre.py

python data_save.py 100

Train the model:

python train_lpkt.py {fold}

For example:

python train_lpkt.py 1 or python train_lpkt.py 2

Test the trained the model on the test set:

python test.py {model_name}

Correction

There is a mistake in the KDD conference paper. Figure. 4 should be results on dataset ASSIST2012

The experimental results would be better than our original paper, as we have optimized the data proprocessing