Matching business process models using positional passage-based language models

Business operations are often documented by business process models. Use cases such as system validation and process harmonization require the identification of correspondences between activities, which is supported by matching techniques that cope with textual heterogeneity and differences in model granularity. In this paper, we present a matching technique that is tailored towards models featuring textual descriptions of activities. We exploit these descriptions using ideas from language modelling. Experiments with real-world process models reveal that our technique increases recall by up to factor five, largely without compromising precision, compared to existing approaches.