Graph-based approaches for rumor detection in social networks: a systematic review

Abstract

Increased public anxiety and fear, disrupted decision-making, social instability, and other significant societal challenges are the results of the rapid spread of rumors on social media platforms. The unique characteristics of these platforms contribute to the rapid spread of both verified and unverified information. These pressing issues highlight the need to develop advanced technologies for early detection and prevention of rumors. This paper presents a systematic review of graph-based approaches for rumor detection in social networks, analyzing 53 studies published between 2018 and 2025. The selected studies are comprehensively reviewed with a focus on graph models and the integration of propagation structure, social, temporal, and content features, which enhances detection accuracy. This review critically evaluates the effectiveness of various methods, highlighting their strengths, limitations, and key challenges. The key contributions of this paper include: (i) an in-depth analysis of current graph-based rumor detection approaches (ii) a categorization of graph models and feature extraction strategies, (iii) the identification of major challenges and research gaps, and (iv) recommendations for future research to develop scalable, robust, and accurate early rumor detection systems. The findings of this study provide valuable insights for researchers aiming to advance the state-of-the-art in fighting misinformation on social networks.