TiSASRec.paddle

A PaddlePaddle implementation of Time Interval Aware Self-Attentive Sequential Recommendation.

Introduction

论文：Time Interval Aware Self-Attentive Sequential Recommendation

Results

Requirement

• Python >= 3
• PaddlePaddle >= 2.0.0
• see requirements.txt

Dataset

MovieLens-1m (max_len = 50)

Usage

Train

bash ./script/train.sh

模型在 200 epochs 左右收敛，日志见 nohup.out。

Test

bash ./script/eval.sh

可以得到如下结果：

References

@inproceedings{li2020time,
  title={Time Interval Aware Self-Attention for Sequential Recommendation},
  author={Li, Jiacheng and Wang, Yujie and McAuley, Julian},
  booktitle={Proceedings of the 13th International Conference on Web Search and Data Mining},
  pages={322--330},
  year={2020}
}

• https://github.com/JiachengLi1995/TiSASRec
• https://github.com/pmixer/TiSASRec.pytorch
• https://github.com/paddorch/SASRec.paddle