Niemann, MichaelMichaelNiemannSiebenhaar, MelanieMelanieSiebenhaarSchulte, StefanStefanSchulteSteinmetz, RalfRalfSteinmetz2022-03-102022-03-102012-02Computers in Industry 63 (2) : 168-180 (2012-02)http://hdl.handle.net/11420/11960Although increasingly IT-supported, effective techniques for process model retrieval and identification of process model differences or changes - needed for a variety of management and conformance purposes - are still challenging problems in business process management. Performing automated process comparison and finding relevant reference processes are not routine procedures for today's operational process repositories. Efficient combinations of similarity measures for various process model characteristics can still improve the performance of process comparison and retrieval. The approach at hand introduces the concept of related cluster pairs, parameterises it with semantic, string-based, and novel hybrid metrics for comparing process models, and defines a novel similarity notion for process model retrieval. Evaluations with process data from the SAP reference model show that our approach outperforms current related work and established text search engines.en0166-361520122168180Process model similarity measureProcess modelsRelated cluster pairsString-based, semantic, and hybrid word similarity metricsAllgemeines, WissenschaftJournal Article10.1016/j.compind.2011.11.002Other