همایش بین المللی هوش مصنوعی

صفحه اصلی / دومین همایش بین المللی هوش مصنوعی

A Comprehensive Review of Big Data Analytics Approaches in Social Networks Using Graph Neural Networks: Methods, Challenges, and Future Directions

نویسندگان :

Mohammad Javad Khedmati Yavabd¹ Hussein Saberi² Mohammad Aali Javadzade³

1- Imam Hussein (AS) Comprehensive University 2- Imam Hussein (AS) Comprehensive University 3- Imam Hussein (AS) Comprehensive University

کلمات کلیدی :

Graph Neural Networks،Big Data،Social Networks, Graph Analytics،Deep Learning،Natural Language Processing،Model Comparison،Algorithmic Challenges

چکیده :

The rapid growth of social media platforms generates massive amounts of complex and interdependent data that traditional machine learning methods cannot analyze effectively. Because this type of data is non-Euclidean and relational in nature, it requires more advanced models that can directly exploit graph structures. GNNs have emerged as strong tools in extracting meaningful structural and behavioral patterns within large-scale social networks by leveraging message-passing and neighborhood aggregation mechanisms. This study presents a systematic and comprehensive review of foundational concepts, major GNN architectures, and their principal applications in social network big data analytics. It also presents a detailed, quantitative comparison of state-of-the-art methods for fundamental tasks such as node classification, link prediction, community detection, sentiment analysis, and fake news detection. In addition, the study investigates some critical challenges faced, such as scalability, noise and incomplete data, privacy issues, and limited generalizability, and discusses their algorithmic implications. Finally, the work points out some emerging research directions: the integration of GNNs with Large Language Models, dynamic and temporal graph learning, federated graph learning, and improved model explainability. This review aims to present researchers with a consolidated updated view of the advances, limitations, and promising future directions of GNN-based analysis in large-scale social networks.