outlook-code / AEGCN

Code for AEGCN

AEGCN

This is our Tensorflow implementation for AEGCN. We will continue format this code.

Environment Requirement

tensorflow-gpu 1.13.1 or 1.12.0

numpy 1.17.0

scipy 1.1.0

pandas 0.25.1

python 3.7.0 or 3.6.2

Examples of Tianchi

missing rate = n%, e.g, n=99, 0%<=n%<100.0%

parameters in parser.py

• Command

python AEGCN.py --dataset tianchi_58 --regs [1e-4] --embed_size 64 --layer_size [64,64,64] --lr 0.001 --batch_size 2048 --epoch 300 --unlabel_rate=0.99 --mtl_dropout=0.2 --lamda2 0.01 --pre_dim2 64

Dataset

We provide the Tiachi dataset used in paper.

Format of Tianchi: user_id,item_id,olditemcategory,itemcategory

Every line is an interaction. itemcategory is the item attribute used in paper (top 300 brands)