With Anchors or Not: Fairness-aware Truss-based Community Search on Attributed Graphs

Xinrui Wang, Zilong Liu, Shixin Ye, Xin Huang, Hong Gao, Xiuzhen Cheng, Dongxiao Yu

Research output: Chapter in book/report/conference proceedingConference proceedingpeer-review

Abstract

Community search, which finds cohesive subgraphs containing given query vertices, has attracted much attention in decades. On attributed graphs, when considering the fairness of members' attributes in a community, the cohesiveness constraint of a clique is too strong, which often causes no fair clique based communities can be found. Thus, in this paper, we use the k-truss model, which is a relaxation of the clique but whose members have large engagement and high tie strength, to describe fair communities, namely fair k-truss communities (FTC) and anchored fair k-truss communities (AFTC, using anchored vertices to help satisfying the fairness constraint). We formulate the FTC and AFTC search problems to find the FTC or AFTC containing a given query vertex q which has the largest k and the smallest diameter. We prove the hardness of both problems. We develop several greedy algorithms and acceleration strategies to solve FTC and AFTC search problems. Experiments on 8 real-world networks show the significance of our FTC and AFTC models, and high performance of our algorithms and acceleration strategies.
Original languageEnglish
Title of host publicationProceedings of the 41st IEEE International Conference on Data Engineering, ICDE 2025
PublisherIEEE
Pages3971-3983
Number of pages13
DOIs
Publication statusPublished - 20 May 2025
Event41st IEEE International Conference on Data Engineering, ICDE 2025 - The Hong Kong Polytechnic University, Hong Kong, China
Duration: 19 May 202523 May 2025
https://ieee-icde.org/2025/
https://ieee-icde.org/2025/research-papers/
https://www.computer.org/csdl/proceedings/icde/2025/26FZy3xczFS

Publication series

NameInternational Conference on Data Engineering

Conference

Conference41st IEEE International Conference on Data Engineering, ICDE 2025
Abbreviated titleICDE 2025
Country/TerritoryChina
CityHong Kong
Period19/05/2523/05/25
Internet address

Fingerprint

Dive into the research topics of 'With Anchors or Not: Fairness-aware Truss-based Community Search on Attributed Graphs'. Together they form a unique fingerprint.

Cite this