Parallel Knowledge Enhancement based Framework for Multi-behavior Recommendation

This repository contains TensorFlow codes and datasets for the paper.

Environment

The code has been tested running under Python 3.6.15. The required packages are as follows:

• nvidia-tensorflow == 1.15.4+nv20.10
• tensorflow-determinism == 0.3.0
• numpy == 1.19.5
• scipy == 1.7.3

Datasets

We utilized three datasets to evaluate PKEF: Beibei, Taobao, and Tmall Contest. The purchase behavior is taken as the target behavior for all datasets. The last target behavior for the test users are left out to compose the testing set. We filtered out users and items with too few interactions.

Just Run It！

• Beibei

python PKEF_final.py --data beibei

• Taobao

python PKEF_final.py --data taobao

• Tmall

python PKEF_final.py --data tmall --gnn_layer "[4, 1, 1, 1]" --coefficient "[0.0/6, 4.0/6, 0.0/6, 2.0/6]"