https://github.com/danison2/MLC-code

This code is based on the paper: A Nonenegative Matrix Factorization Approach for Multiple Local Community Detection published in the ASONAM conference in 2018.
https://github.com/danison2/MLC-code

Last synced: about 1 month ago
JSON representation

This code is based on the paper: A Nonenegative Matrix Factorization Approach for Multiple Local Community Detection published in the ASONAM conference in 2018.

• Host: GitHub
• URL: https://github.com/danison2/MLC-code
• Owner: danison2
• Created: 2019-10-09T11:01:15.000Z (about 5 years ago)
• Default Branch: master
• Last Pushed: 2024-06-17T23:07:12.000Z (6 months ago)
• Last Synced: 2024-08-01T17:36:01.388Z (4 months ago)
• Language: Python
• Homepage:
• Size: 8.54 MB
• Stars: 3
• Watchers: 1
• Forks: 0
• Open Issues: 2
• Metadata Files:
  • Readme: readme.txt

Awesome Lists containing this project

• awesome-community-detection - [Python Reference

README

        
This code is based on the paper: A Nonenegative Matrix Factorization Approach for Multiple Local Community Detection published in the ASONAM conference in 2018.

To run  the code with the sample Amazon network:

(1) with cmd go to the code folder

(2) pip install -r requirements.txt

(3) go to MLC folder

(4)python MLC.py

(5) go to MLC-code folder, you will find the conductance results in the Cond folder and the F1 results in the F1 folder.

NOTE: 

(a) graphA is Amazon while graphD is DBLP.

(b) To run this code on a different graph, change the following variables:

	graphFiles=['graphA.txt'] 	#Amazon

	communityFile='newComA.txt' 	#cleaned ground-truth communities with duplicates removed

        seedsFiles=['seedsA3.txt'] 	#seeds that belong to three communities

        delimiter = "\t" 		#delimiter of the graph's edge list. For some graphs, it is just blank space " "

The data folder contains other sample graphs and seeds, and their ground-truth communities. 

Karate club is not included in the graphs folder as it can be generated using: G = nx.karate_club_graph()

Citation (BibTex format):

@article{Kamuhanda2018ANM,

  title={A Nonnegative Matrix Factorization Approach for Multiple Local Community Detection},

  author={Dany Kamuhanda and Kun He},

  journal={2018 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining (ASONAM)},

  year={2018},

  pages={642-649}

}