CMiner: An Algorithm to Discover Frequent Structures in Conceptual Models

Proc. of the 29th Int. Conf. on Extending Database Technology (EDBT). 2026.

The analysis of conceptual models to find recurrent structures and patterns is a lively research area, aimed at finding good or bad modeling practices or, more generally, recurrent phenomena represented within the model. In the last few years, several approaches have been explored using automated techniques to support such a discovery process. However, due to the complex structure of the graphs encoding conceptual models and the challenging nature of the discovery task, the available techniques can still be largely enhanced. This paper presents a novel Frequent Subgraph Mining (FSM) algorithm called CMiner, designed specifically for discovering recurrent structures in conceptual models, which offers a guided, performance-oriented approach to mining semantically rich graphs. The proposed solution is inspired by established methods but is tailored to address specific needs arising in the conceptual modeling context. It supports heuristic tasks, enhances the discovery of recurring structures, and thus can serve as a key ally in pattern identification. Besides providing a detailed overview of CMiner, which is freely available, we validate it as a useful support tool for pattern discovery and compare it with state-of-the-art solutions.

@inproceedings{EDBT-2026,
   title = "CMiner:  An Algorithm to Discover Frequent Structures in
Conceptual Models",
   year = "2026",
   author = "Simone Avellino and Emanuele Valore and Giovanni Micale and
Antonio Di Maria and Mattia Fumagalli and Tiago Prince Sales and Alfredo
Pulvirenti and Diego Calvanese",
   booktitle = "Proc. of the 29th Int. Conf. on Extending Database Technology
(EDBT)",
   pages = "345--358",
   publisher = "OpenProceedings.org",
   doi = "10.48786/EDBT.2026.28",
}