This repository is a Tensorflow implementation of HierTCN for Dynamic Recommender Systems.

Jiaxuan You, Yichen Wang, Aditya Pal, Pong Eksombatchai, Chuck Rosenberg, Jure Leskovec, Hierarchical Temporal Convolutional Networks for Dynamic Recommender Systems, The Web Conference 2019 (WWW-2019).

Tensorflow (tested on 1.6.0), numpy, pandas, pickle

Download XING dataset, only interactions.csv is needed. Create dataset/ folder, then copy interactions.csv to that folder.

If XING dataset is not available, you can email Jiaxuan You for the dataset for solely research use.

python run_xing.py

Each run will generate a folder in ./run, which can be accessed via tensorboard. A summarized text file is also generated.