Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

GCTAM: Global and Contextual Truncated Affinity Combined Maximization Model For Unsupervised Graph Anomaly Detection

About

Anomalies often occur in real-world information networks/graphs, such as malevolent users, malicious comments, banned users, and fake news in social graphs. The latest graph anomaly detection methods use a novel mechanism called truncated affinity maximization (TAM) to detect anomaly nodes without using any label information and achieve impressive results. TAM maximizes the affinities among the normal nodes while truncating the affinities of the anomalous nodes to identify the anomalies. However, existing TAM-based methods truncate suspicious nodes according to a rigid threshold that ignores the specificity and high-order affinities of different nodes. This inevitably causes inefficient truncations from both normal and anomalous nodes, limiting the effectiveness of anomaly detection. To this end, this paper proposes a novel truncation model combining contextual and global affinity to truncate the anomalous nodes. The core idea of the work is to use contextual truncation to decrease the affinity of anomalous nodes, while global truncation increases the affinity of normal nodes. Extensive experiments on massive real-world datasets show that our method surpasses peer methods in most graph anomaly detection tasks. In highlights, compared with previous state-of-the-art methods, the proposed method has +15\% $\sim$ +20\% improvements in two famous real-world datasets, Amazon and YelpChi. Notably, our method works well in large datasets, Amazin-all and YelpChi-all, and achieves the best results, while most previous models cannot complete the tasks.

Xiong Zhang, Hong Peng, Zhenli He, Cheng Xie, Xin Jin, Hua Jiang• 2026

Related benchmarks

TaskDatasetResultRank
Graph Anomaly DetectionWeibo (test)
AUPRC16.87
39
Graph Anomaly DetectionFacebook (test)--
32
Graph Anomaly DetectionReddit (test)--
32
Graph Anomaly DetectionCiteseer (test)
AUPRC0.1029
30
Anomaly DetectionFacebook zero-shot (test)
AUROC68.61
19
Anomaly DetectionACM zero-shot (test)
AUROC80.14
19
Anomaly Detectionphoto zero-shot (test)
AUROC63.12
19
Anomaly DetectionAnomaly Detection Challenge Zero-Shot blind (test)
AUROC74.12
19
Anomaly DetectionCiteSeer zero-shot (test)
AUROC70.31
19
Anomaly DetectionReddit zero-shot (test)
AUROC (zero-shot)59.32
19
Showing 10 of 20 rows

Other info

Follow for update