Arya, Aikta; Pandey, Pradumn Kumar; Ganguly, Niloy; Derr, Tyler. (2026).听.听ACM Transactions on Knowledge Discovery from Data, 20(5), 76.听
Online social networks聽have transformed how people share information, driving the development of technologies such as recommendation systems, community detection, and tools that predict relationships between users. Developing and testing these methods requires high-quality network data, but real-world datasets are often limited by privacy concerns, availability, or incomplete information. One solution is to generate聽synthetic signed networks鈥攁rtificial datasets that model both positive and negative relationships, such as trust and distrust between users. This review surveys current methods for reconstructing these signed networks, compares their performance through experimental evaluations, and examines their applications in social network analysis and machine learning. The authors also discuss the strengths and limitations of existing approaches, identify key challenges and open research questions, and outline opportunities for future work. By bringing together the current state of the field, this survey provides a resource for researchers working to develop more realistic synthetic network models and improve algorithms that rely on signed network data.
Fig 1
Services of social networks and its underlying frontend and backend algorithms