GrFrauder: A Novel Unsupervised Clustering Algorithm for Identification Group Spam Reviewers

Abstract

As e-commerce has expanded, people's lives now include some aspect of online buying, because buyers frequently use online product reviews to make purchasing decisions. Merchants frequently collaborate with review spammers to write spam reviews that promote or demote selected items. Spammers who work in groups, in particular, are more dangerous than individual attacks. Previous studies provided various frequent item mining and graphbased techniques to detect such spammer groups. In this paper, we recommend a technique referred to as GrFrauder (Group Fraud detection) method to detect online spam reviewer groups with an unsupervised manner. Our technology identifies spammer candidate groups initially based on product -product review graph and collaboration among reviewers constructed with several behavioral patterns. It then embeds reviewers into an embedding space and calculates spam score for every group; with higher spam scores the model generates ranks for each group. Studies using four real-world datasets reveal that GrFrauder outperforms numerous state-of-the-art baselines in terms of precision and is able to identify more high-quality spammer groups.

Multi-granularity Signal Processing Method for Digital Twin Power Grids via Graph Representation Learning

“…Recently, numerous deep learning techniques have emerged to facilitate clustering based on topology, attributes, behaviors, and other aspects [10][11][12]. Algorithms like DeepWalk [13], Node2Vec [14], and LINE [15] have established themselves as classical methods in complex network representation learning, effectively addressing the challenge of preserving local topology. Such as SDNE [16] and GCN [17], these studies achieve clustering by mapping individual nodes to different levels of granularity, i.e., by considering the topology of nodes, as in MNRL [18], by considering the topology between nodes and the properties of neighboring nodes.…”