/PPCF

PPCF: Privacy-Preserving Collaborative Filtering for Web Service Recommendation

Primary LanguagePythonMIT LicenseMIT

PPCF

This repository maintains privacy-preserving QoS prediction approaches for Web service recommendation.

Read more information from our paper:

Jieming Zhu, Pinjia He, Zibin Zheng, and Michael R. Lyu, "A Privacy-Preserving QoS Prediction Framework for Web Service Recommendation," in Proc. of IEEE ICWS, 2015. [Paper][Project page]

Dependencies

  • Python 2.7
  • numpy
  • scipy
  • cython

Usage

The benchmark is implemented in Python. For efficiency purpose, the core algorithms are written as Python extension using C++.

  1. Download the package: https://github.com/wsdream/PPCF/archive/master.zip,

    or use Git: git clone https://github.com/wsdream/PPCF.git.

  2. Download WS-DREAM datasets to the data folder.

  3. Build extension modules based on Cython

    $ cd PPCF/
$ python setup.py build_ext --inplace

  4. Run the demo scripts

    $ cd PPCF/demo/P-PMF
$ python run_rt.py
$ python run_tp.py

  5. Check the evaluation results in "result/" directory. Note that we have already provided the results of 20 random runs in the result directory for your quick reference.

Feedback

For bugs or feedback, please post to our issue page.

License

The MIT License (MIT)