Community-Aware Social Community Recommendation

2025-08-07

AI Summary

Existing social recommendation models exhibit limited performance in community recommendation tasks, primarily due to their neglect of the inherent dynamic evolution of communities and the rich global–local structural patterns present in social networks. To address this, we propose CASO—a novel community-aware social recommendation model that, for the first time, unifies social modularity maximization, social cohesion aggregation, and collaborative filtering within a single framework. CASO incorporates a mutual exclusion mechanism between social and collaborative signals to enhance community awareness and employs a graph encoder to jointly learn user and community embeddings. It optimizes multiple objectives—including modularity, cohesion, collaborative filtering, and community detection—via a composite loss function. Extensive experiments on six real-world social networks demonstrate that CASO consistently outperforms nine state-of-the-art baselines, validating its effectiveness and superiority in community recommendation.

Technology Category

• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Application Domains: Humanities & Computational Social Science

Application Category

• User Modeling, Personalization and Recommendation: Social recommender systems and personalization for the social Web
• Social Networks and Social Media: Computational social science
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Social recommendation, which seeks to leverage social ties among users to alleviate the sparsity issue of user-item interactions, has emerged as a popular technique for elevating personalized services in recommender systems. Despite being effective, existing social recommendation models are mainly devised for recommending regular items such as blogs, images, and products, and largely fail for community recommendations due to overlooking the unique characteristics of communities. Distinctly, communities are constituted by individuals, who present high dynamicity and relate to rich structural patterns in social networks. To our knowledge, limited research has been devoted to comprehensively exploiting this information for recommending communities. To bridge this gap, this paper presents CASO, a novel and effective model specially designed for social community recommendation. Under the hood, CASO harnesses three carefully-crafted encoders for user embedding, wherein two of them extract community-related global and local structures from the social network via social modularity maximization and social closeness aggregation, while the third one captures user preferences using collaborative filtering with observed user-community affiliations. To further eliminate feature redundancy therein, we introduce a mutual exclusion between social and collaborative signals. Finally, CASO includes a community detection loss in the model optimization, thereby producing community-aware embeddings for communities. Our extensive experiments evaluating CASO against nine strong baselines on six real-world social networks demonstrate its consistent and remarkable superiority over the state of the art in terms of community recommendation performance.

Problem

• Recommending social communities by leveraging user social ties
• Addressing unique community dynamics and structural patterns
• Reducing feature redundancy between social and collaborative signals

Innovation

• Uses three encoders for user embedding
• Applies social modularity and closeness aggregation
• Incorporates community detection loss optimization