A reimplementation of SRGNN.
(WARNING: The computation of session embedding only uses embedding W.R.T. nodes in a session graph while in the paper, the calculation based on the whole session sequence, which means they may calculate re-occur items as many times as they occue.)
Original code from here. Original paper.
Borrow the data preprocessing from original repo, including diginetica
and yoochoose
.
Using PyTorch 1.0, TensorboardX and PyTorch-Geometric.
-
Follow the steps in original code repo to get
train.txt
andtest.txt
for every dataset. -
Put both
txt
file in theraw
folder W.R.T. different datasets.
cd src
python main.py --dataset=diginetica
If you use make advantage of the SR-GNN model in your research, please cite the following:
@inproceedings{Wu:2019vb,
author = {Wu, Shu and Tang, Yuyuan and Zhu, Yanqiao and Wang, Liang and Xie, Xing and Tan, Tieniu},
title = {Session-based Recommendation with Graph Neural Networks},
booktitle = {Proceedings of The Twenty-Third AAAI Conference on Artificial Intelligence},
series = {AAAI '19},
year = {2019},
url = {http://arxiv.org/abs/1811.00855}
}