Abstract | ||
---|---|---|
It is common for the observed behavior of a business process to differ from the behavior captured in its corresponding model, as workers devise workarounds to handle special circumstances, which over time become part of the norm. Process model repair methods help modelers to realign their models with the observed behavior as recorded in an event log. Given a process model and an event log, these methods produce a new process model that more closely matches the log, while resembling the original model as close as possible. Existing repair methods identify points in the process where the log deviates from the model, and fix these deviations by adding behavior to the model locally. In their quest for automation, these methods often add too much behavior to the model, so that the repaired model grossly over-generalizes the log. This paper advocates for an interactive and incremental approach to process model repair, where differences between the model and the log are visually displayed to the user, and the user repairs each difference manually based on the provided visual guidance. An empirical evaluation shows that the proposed method leads to repaired models that avoid the over-generalization pitfall of state-of-the-art automated repair methods. |
Year | Venue | Field |
---|---|---|
2017 | OTM Conferences | Visual guidance,Workaround,Software engineering,Business process,Visualization,Automation,Artificial intelligence,Business process modeling,Conformance checking,Engineering,Machine learning |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 5 |
Name | Order | Citations | PageRank |
---|---|---|---|
Abel Armas-Cervantes | 1 | 34 | 6.37 |
Nick van Beest | 2 | 18 | 6.77 |
marcello la rosa | 3 | 1402 | 81.70 |
Marlon Dumas | 4 | 5742 | 371.10 |
Luciano Garcia-Banuelos | 5 | 870 | 43.84 |