IMP_GCN

This is our implementation for the paper:

Fan Liu, Zhiyong Cheng*, Lei Zhu, Zan Gao and Liqiang Nie*. Interest-aware Message-Passing GCN for Recommendation. The Web Conference 2021 (WWW'21), Ljubljana, Slovenia, 2021 (“*”= Corresponding author)

Environment Settings

• Tensorflow-gpu version: 1.3.0

Example to run the codes.

gowalla

Run IMP_GCN.py

python IMP_GCN.py --dataset gowalla  --regs [1e-4] --embed_size 64 --layer_size [64,64,64,64,64,64] --lr 0.001 --batch_size 2048 --epoch 2000 --groups 3 --Ks [20,10] --gpu_id 0