Encoding methods affect the performance of process mining tasks but little work in the literature focused on quantifying their impact. In this paper, we compare 10 different encoding methods from three different families (trace replay and alignment, graph embeddings, and word embeddings) using measures to evaluate the overlaps in the feature space, the accuracy obtained, and the computational resources (time) consumed with a classification task. Across hundreds of event logs representing four variations of five scenarios and five anomalies, it was possible to identify the edge2vec method as the most accurate and effective in reducing class overlapping in the feature space.
Evaluating Trace Encoding Methods in Process Mining / S. Barbon Junior, P. Ceravolo, E. Damiani, G. Marques Tavares (LECTURE NOTES IN ARTIFICIAL INTELLIGENCE). - In: From Data to Models and Back / [a cura di] J. Bowles, G. Broccia, M. Nanni. - [s.l] : Springer, 2021. - ISBN 9783030706494. - pp. 174-189 (( Intervento presentato al 9. convegno DataMod nel 2020.
|Titolo:||Evaluating Trace Encoding Methods in Process Mining|
MARQUES TAVARES, GABRIEL (Corresponding)
|Parole Chiave:||Trace encoding; Word embeddings; Graph embeddings; Classification; Process Mining|
|Settore Scientifico Disciplinare:||Settore INF/01 - Informatica|
|Data di pubblicazione:||2021|
|Digital Object Identifier (DOI):||http://dx.doi.org/10.1007/978-3-030-70650-0_11|
|Tipologia:||Book Part (author)|
|Appare nelle tipologie:||03 - Contributo in volume|