HIN-RNN: A Graph Representation Learning Neural Network for Fraudster Group Detection With No Handcrafted Features

Shehnepoor, Saeedreza; Togneri, Roberto; Liu, Wei; Bennamoun, Mohammed

doi:10.1109/tnnls.2021.3123876

Cited by 13 publications

(24 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the approaches in both categories suffer from significant drawbacks. First, the most recent approaches (Ji et al [11], Zhang et al [12], and Shehnepoor et al [15]) overlooked the temporal nature of fraudster groups, as illustrated in Fig. 1.…”

Section: Introductionmentioning

confidence: 99%

“…Such outliers; i.e., fraudster(s) in a genuine group or genuine reviewer(s) in a fraudster group; pollute the group, creating another level of challenges for fraudster group detection. The most recent approach of fraudster group detection by Shehenepoor et al [15] excluded such outliers by removing the users with the least connection with other reviewers. However, such an approach overlooks the importance of the joint representation in covering different aspects of a reviewer's activities in a social platform.…”

Section: Introductionmentioning

confidence: 99%

“…A, during T 1 , reviewer r 1 and r 2 have a co-review relationship whereas during T N , r 3 and r 4 have co-reviewed. We take advantage of the HIN-RNN [15] to model the co-review relation between reviewers, as it has been proven to be effective in utilizing spatial relations between reviewers. HIN-RNN [15] is an RNN used to model the relation between nodes in a graph considering the nodes' heterogeneity.…”

Section: Introductionmentioning

confidence: 99%

“…We take advantage of the HIN-RNN [15] to model the co-review relation between reviewers, as it has been proven to be effective in utilizing spatial relations between reviewers. HIN-RNN [15] is an RNN used to model the relation between nodes in a graph considering the nodes' heterogeneity. The heterogeneity is obtained through a fine-tuning of the semantics extracted using word embedding techniques based on the reviewer's type (fraudster/genuine).…”

Section: Introductionmentioning

confidence: 99%

“…). The relations between reviewers in groups for different time windows are then refined with the reviewers' representation by the HIN-RNN [15] to output the spatial relations. The HIN-RNN realizes a possible spatial correlation between r 2 and r 3 in the 1 st and N th time window.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Spatio-Temporal Graph Representation Learning for Fraudster Group Detection

Shehnepoor,

Togneri,

Liu

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Motivated by potential financial gain, companies may hire fraudster groups to write fake reviews to either demote competitors or promote their own businesses. Such groups are considerably more successful in misleading customers, as people are more likely to be influenced by the opinion of a large group. To detect such groups, a common model is to represent fraudster groups static networks, consequently overlooking the longitudinal behavior of a reviewer thus the dynamics of co-review relations among reviewers in a group. Hence, these approaches are incapable of excluding outlier reviewers, which are fraudsters intentionally camouflaging themselves in a group and genuine reviewers happen to co-review in fraudster groups. To address this issue, in this work, we propose to first capitalise on the effectiveness of the HIN-RNN in both reviewers' representation learning while capturing the collaboration between reviewers, we first utilize the HIN-RNN to model the co-review relations of reviewers in a group in a fixed time window of 28 days. We refer to this as spatial relation learning representation to signify the generalisability of this work to other networked scenarios. Then we use an RNN on the spatial relations to predict the spatiotemporal relations of reviewers in the group. In the third step, a Graph Convolution Network (GCN) refines the reviewers' vector representations using these predicted relations. These refined representations are then used to remove outlier reviewers. The average of the remaining reviewers' representation is then fed to a simple fully connected layer to predict if the group is a fraudster group or not. Exhaustive experiments of the proposed approach showed a 5% (4%), 12% (5%), 12% (5%) improvement over three of the most recent approaches on precision, recall, and F1-value over the Yelp (Amazon) dataset, respectively.

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%