---

Recurrent Tensor Factorization

This is the implementation of our paper:

Yiwen Zhang, Chunhui Yin, Zhihui Lu*, Dengcheng Yan, Meikang Qiu, Qifeng Tang. Recurrent Tensor Factorization for time-aware Service Recommendation, Applied Soft Computing 85 (2019) 105762. (SCI)

Author: Chun-hui Yin

Affiliate: Big Data and Cloud Service Lab, Anhui University

Last updated: 2019/10/05

Please cite our paper if you use our codes. Thanks!

Environment Requirement

This code can be run at following requirement but not limit to:

• python = 3.6.6
• tensorflow-gpu = 1.7.0
• keras = 2.0.9
• pandas = 0.23.4
• numpy = 1.14.0
• scikit-learn = 0.21
• other installation dependencies required above

Example of Usage

>>>python RTF.py

>>>python GTF.py

>>>python PGRU.py

Dataset

• To simulate the real-world situation, we sparse the original matrix at 4 densities and generate instances for training
• Here we provide the preprocessed real-world dataset WS-Dream (dataset#2)
• The original WS-DREAM dataset can be downloaded at InplusLab

Note

• Experiments can be run on multi-core CPUs at 6 densities by turning on parallel mode