Researchers unveil apnet: graph tool screens academic publishing for anomalies

A new paper from Robert Šamárek and Radek Martinek at VSB-Technical University of Ostrava introduces apnet, an open-source graph analytics library for auditing academic publishing integrity. Instead of relying on simplistic metrics like the h-index or Journal Impact Factor, apnet models the entire scholarly ecosystem as a heterogeneous graph with seven node types (works, authors, institutions, journals, topics, etc.) and seven edge types (citations, co-authorship, affiliation, publication venue). The library computes structural metrics, performs community detection, and runs three screening detectors that flag anomalous publishing patterns—such as dense co-authorship cliques, locally closed citation loops, and thematically isolated venues. The researchers deliberately avoid binary classification; detectors return ranked candidates with explicit structural evidence for human assessment.

Evaluation on VSB-TUO's institutional corpus (2020–2025) showed that community detection recovers real research groups and centralities identify cross-disciplinary bridges. When testing on a venue-centric corpus with external ground truth (journals delisted by Scopus and DOAJ), a naive case-control design yielded strong but spurious detectors due to a prominence confound. After size-matching, the only robust signal was the breadth of disciplinary scope (AUC 0.70). The paper also shows that PageRank over the journal citation network tracks a JIF proxy while being an order of magnitude more resistant to citation gaming than count-based indicators. The apnet library, released as open source with a CLI and web interface, processes the analysis on commodity hardware in minutes, making large-scale integrity screening accessible to institutions and publishers.