menghuaa / Link_prediction

Link prediction for network using node2vec on the database of ca-AstroPh, Facebook, Protein-Protein Interaction, Wikipedia

Link_prediction

Link prediction for networks using node2vec

Datasets

Ca-AstroPh, Facebook, Protein-Protein Interaction, Blog Catalog, Wikipedia

Usage

For Node2Vec mode, it's the default, no argument is needed:

python link_prediction.py --r 3 --f dataset_name

For deepwalk mode (after installing it):

python link_prediction.py --m deepwalk --r 3 --f dataset_name

List of all command line arguments

• --r : number of executions (default 1)
• --m : node embedding mode (default 'Node2Vec')
• --f : input dataset file name, an edge list in .txt format (default 'ca-AstroPh.txt')
• --l : log file name, where results are stored (default 'log.txt')

Run

python link_prediction.py --h

to see full list in prompt