TY - GEN
T1 - From Semantically Abstracted Traces to Process Mining and Process Model Comparison
AU - Leonardi, Giorgio
AU - Striani, Manuel
AU - Quaglini, Silvana
AU - Cavallini, Anna
AU - Montani, Stefania
N1 - Publisher Copyright:
© 2018, Springer Nature Switzerland AG.
PY - 2018
Y1 - 2018
N2 - Process model comparison can be exploited to assess the quality of organizational procedures, to identify non-conformances with respect to given standards, and to highlight critical situations. Sometimes, however, it is difficult to make sense of large and complex process models, while a more abstract view of the process would be sufficient for the comparison task. In this paper, we show how process traces, abstracted on the basis of domain knowledge, can be provided as an input to process mining, and how abstract models (i.e., models mined from abstracted traces) can then be compared and ranked, by adopting a similarity metric able to take into account penalties collected during the abstraction phase. The overall framework has been tested in the field of stroke management, where we were able to rank abstract process models more similarly to the ordering provided by a domain expert, with respect to what could be obtained when working on non-abstract ones.
AB - Process model comparison can be exploited to assess the quality of organizational procedures, to identify non-conformances with respect to given standards, and to highlight critical situations. Sometimes, however, it is difficult to make sense of large and complex process models, while a more abstract view of the process would be sufficient for the comparison task. In this paper, we show how process traces, abstracted on the basis of domain knowledge, can be provided as an input to process mining, and how abstract models (i.e., models mined from abstracted traces) can then be compared and ranked, by adopting a similarity metric able to take into account penalties collected during the abstraction phase. The overall framework has been tested in the field of stroke management, where we were able to rank abstract process models more similarly to the ordering provided by a domain expert, with respect to what could be obtained when working on non-abstract ones.
UR - http://www.scopus.com/inward/record.url?scp=85057402060&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-03840-3_4
DO - 10.1007/978-3-030-03840-3_4
M3 - Conference contribution
AN - SCOPUS:85057402060
SN - 9783030038397
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 47
EP - 59
BT - AI*IA 2018 – Advances in Artificial Intelligence - 17th International Conference of the Italian Association for Artificial Intelligence, Proceedings
A2 - Ghidini, Chiara
A2 - Traverso, Paolo
A2 - Magnini, Bernardo
A2 - Passerini, Andrea
PB - Springer Verlag
T2 - 17th Conference of the Italian Association for Artificial Intelligence, AI*IA 2018
Y2 - 20 November 2018 through 23 November 2018
ER -