A comprehensive review: Anomaly detection techniques on social networking and its applications

Abstract

Due to large part in the expansion of Web 2.0 and the Internet over the last ten years, interest in online social networks has increased dramatically. These websites are among the most widely used worldwide, with applications in telemarketing, healthcare, education, and entertainment that touch almost every facet of daily life. Unfortunately, criminals have made these their primary targets when attempting to commit crimes and harm other users. It is possible to identify the strange actions of these consumers by employing anomaly detection techniques. Anomaly detection in social networks involves identifying unusual and unexpected user behavior by analyzing the hidden patterns within the networks. This task is particularly challenging because the interaction patterns of these users differ significantly from those of regular network users. Although numerous anomaly detection techniques have been designed for various problematic situations, this area is still comparatively new and quickly evolving. Consequently, a structured analysis of the research conducted in anomaly detection within social networks is increasingly necessary. This paper provides an in-depth analysis of various approaches to social network anomaly detection. It introduces multiple taxonomy levels that categorize current techniques based on the types of anomalies they detect, the characteristics of the input networks, and the underlying detection methodologies. Additionally, the study explores the open problems and research obstacles in this field and highlights the diverse application contexts where these methods have been employed.