arrow
Return

A robust two-stage algorithm for local community detection

delete2018-07-01
delete63
PRE
AI
X
Xiaoyu Ding
J
Jianpei Zhang *
杨静 (Jing Yang)
DOI:10.1016/j.knosys.2018.04.018delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Local community detection addresses the efficiency problem faced by global community detection. Most existing local community detection algorithms take a seed as an initial community. They extend the community by running a greedy optimization process for a quality function. However, the quality of the detected community depends on the location of the seed. This leads to seed-dependent problem. Besides that, many local community detection algorithms cannot ensure the seed exists in the detected community. This leads to seed-invalid problem. This article proposes a robust two-stage local community detection algorithm (RTLCD) based on core detecting and community extension. To solve the seed-dependent problem, the core detecting stage replaces the seed with the core member of the target community. To solve the seed-invalid problem, the community extension stage takes the detected community core member as an initial community and extends the community based on relation strength. Experimental results on artificial and real-world networks show that RTLCD is more robust to the seed-dependent problem and the seed-invalid problem than earlier state-of-the-art algorithms. In addition, RTLCD has excellent performance in identifying more ground-truth community members. (C) 2018 Elsevier B.V. All rights reserved.
Keywords:
Local community detection
Seed-dependent problem
Seed-invalid problem
Core detecting
Community extension
AI Summary

AI Summary

Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.

Journal

K
Knowledge-Based Systems
IF:
7.6
Papers:
1.2W
Citations:
4.5W

Organization

H
Harbin Engineering University
Scholars:
1.9W
Papers: 1.3W
Citations: 1.3W