anirban-code-to-live / line2vec

Information network edge representation learning using edge-to-vertex dual graphs (a.k.a line graph). In addition to that, an optimisation problem is solved efficiently to generate the edge embeddings.

line2vec

The line2vec algorithm learns continuous representations for edges in any (un)directed, (un)weighted graph. It is an information network edge representation learning using edge-to-vertex dual graphs (a.k.a line graph). In addition to that, an optimisation problem is solved efficiently to generate the edge embeddings.

Environment Set-up

Use python version 2.7

• Clone the repository.
• Navigate to the base directory of line2vec (the download location)
• Create a virtual environment using the following command:
  virtualenv venv
  (If virtualenv package is not installed, please install using pip)
• Activate the environment using:
  source venv/bin/activate
• Install required python modules to run the code.
  pip install -r requirements.txt

Congratulations!! You are now setup to run the code.

Basic Usage

Input

• Look at the sample dataset karate inside data directory. If you want to experiment on different datasets, create a folder with name of your dataset.
• The edgelist formatted file corresponding to the link structure of the information network is required to run and generate the embedding.
• Naming convention for the edgelist file: <dataset_name>.edgelist.
• Note : If you are running the code for the first time, include the scratch flag in the run command.

Example

To run line2vec on Zachary's karate club network, execute the following command, navigate to the src directory inside the project's home path and run the following command:

python main.py --input ../data/karate/karate.edgelist --dataset karate --output ../embed/karate/karate.embed --dimensions=8 --line-graph ../data/karate/karate_line.edgelist --l2v-iter=20 --iter=2

Options

You can check out the other options available to use with line2vec using:
python src/main.py --help

Input

The supported input format is an edgelist:

node1_id_int node2_id_int <weight_float, optional>

The graph is assumed to be undirected and unweighted by default. These options can be changed by setting the appropriate flags.

Output

The output file has n+1 lines for a graph with n edges. The first line has the following format:

num_of_edges dim_of_representation

The next n lines are as follows:

edge_id dim1 dim2 ... dimd

where dim1, ... , dimd is the d-dimensional representation learned by line2vec.