a simplified python implementation for DeepLink DeepLink: A Deep Learning Approach for User Identity Linkage

• python >= 2.7
• tensorflow >= 1.3.0
• numpy >= 1.14.0
• gensim >= 3.4.0

The dataset used in this project are from Aligning users across social networks using network embedding. Due to the privacy concern, we do not provide all the raw data. However, the embedding data and some train data can be found in the data folder.

To run DeepLink, first clone the project to your python IDE (eg:Pycharm), then run the main.py. Our embedding method is introduced in the file embedding.py, which is a random_walk and word2vec implementation .

Note: you need to install the required libs.