This is a fork of the original DeepWalk available here
https://github.com/phanein/deepwalk
Unlike the original DeepWalk, this fork only extracts walks from a list of edges, not including the optimization step used to build graph node embeddings.
To build node embeddings, you can use any suitable optimizer. For example, word2vec:
https://github.com/tmikolov/word2vec
To install this fork, your computer simply needs to have installed Python 3.6 and Pip. If so, use this command
pip install -U "deepwalk @ git+https://github.com/aedera/deepwalk.git"
Once installed, DeepWalk can be run from command line with this command
deepwalk examples/edges.txt 0 2 10 1234
This extracts approximately 2 walks of length 10 from each node for the graph
described in the input file examples/edges.txt
assuming directed edges (0,
use 1 to assume undirected edges). The last parameter (1234) sets a seed for
replication.
The above command outputs the walks to standard output.