SPACED: A Novel Deep Learning Method for Community Detection in Social Networks

Mohammed Tirichine, Nassim Ameur, Younes Boukacem, Hatem M. Abdelmoumen, Hodhaifa Benouaklil, Samy Ghebache, Boualem Hamroune, Malika Bessedik, Fatima Benbouzid Si Tayeb, Riyadh Baghdadi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Community detection is a landmark problem in social network analysis. To address this challenge, we propose SPACED: Spaced Positional Autoencoder for Community Embedding Detection, a deep learning-based approach designed to effectively tackle the complexities of community detection in social networks. SPACED generates neighborhood-aware embeddings of network nodes using an autoencoder architecture. These embeddings are then refined through a mixed learning strategy with generated community centers, making them more community-aware. This approach helps unravel network communities through an appropriate clustering strategy. Experimental evaluations across synthetic and real-world networks, as well as comparisons with state-of-the-art methods, demonstrate the high competitiveness and often superiority of SPACED for community detection while maintaining reasonable time complexities.

Original languageEnglish (US)
Title of host publicationProceedings of the 20th International Conference on Web Information Systems and Technologies, WEBIST 2024
EditorsFrancisco Garcia Penalvo, Karl Aberer, Massimo Marchiori
PublisherScience and Technology Publications, Lda
Pages141-152
Number of pages12
ISBN (Electronic)9789897587184
DOIs
StatePublished - 2024
Event20th International Conference on Web Information Systems and Technologies, WEBIST 2024 - Porto, Portugal
Duration: Nov 17 2024Nov 19 2024

Publication series

NameInternational Conference on Web Information Systems and Technologies, WEBIST - Proceedings
ISSN (Print)2184-3252

Conference

Conference20th International Conference on Web Information Systems and Technologies, WEBIST 2024
Country/TerritoryPortugal
CityPorto
Period11/17/2411/19/24

Keywords

  • Community Detection
  • Community Embedding
  • Deep Learning
  • Node Embedding
  • Social Network

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems

Fingerprint

Dive into the research topics of 'SPACED: A Novel Deep Learning Method for Community Detection in Social Networks'. Together they form a unique fingerprint.

Cite this