Tensorflow implementation of paper
This implementation contains:
- DeepWalk
- GraphSage
- DSPR
- Metapath++
- BiNE
- IntentGC
- IntentGC++
Requirements
- Python 2.7
CopyRight Statement
- These codes should only be used for offline research/study purpose. Any usage in bussiness online production or other purpose is not permitted. Anyone who copy these codes for other purpose should be responsible for any consequences or legal liability.
Owner
- Jun Zhao
Note
- These codes are deployed on public amazon data (which a subdataset provided by amazon) and contains a light version of IntentGC and IntnetGC++ on the amazon data. It should be note that this is a separete implementation which is different from the real version implemented on Taobao. So this piece of codes are not representing the produciton of Alibaba. According to company rule, we can only open entirely independent implementaion of experiments on public amazon data, and this is a personal work for open source contribution and research efforts. For the complete ideas on algorithms, please refer to the paper.
- The category in the code refers to the property in the paper, the term in the code refers to the word in the paper, just name alias.
Contact
- For anyone who have problem or are interested about further detailed implementaion in research purpose, welcome to contact Jun Zhao (The email is on the paper)
thrid party usage
In this implementation, we adopt APIs of tensorflow, numpy, sklearn. Thanks these third party open source platform/tools.
package denpendance
To run these codes, you still need to install a graph computing platform named Eular, which will be officially released by Alimama Engine Team soon. We will add the install path later.
License
License.