The algorithm learns simultaneously from multiple unipartite and bipartite graphs and produces the corresponding embeddings.
- size: The size of the dimention
- negative: The number of the Negative Samples
- samples: Total number of samples (in millions)
- rho: Learning rate
- threads: How many threads it will create
- path: The path to the input files
- Foursquare: https://sites.google.com/site/dbhongzhi/
- Gowalla, Weeplaces: http://www.yongliu.org/datasets/