Fast conformance analysis based on activity log abstraction

    Onderzoeksoutput: Hoofdstuk in Boek/Rapport/CongresprocedureConferentiebijdrageAcademicpeer review

    2 Citaten (Scopus)

    Samenvatting

    Process mining techniques focus on bridging the gap between activity logs and business process management. Process discovery is a sub-field of process mining which uses activity logs in order to discover process models. Some process discovery techniques, such as interactive process discovery and genetic algorithms, rely on the so-called conformance analysis. In such techniques, process models are discovered in an incremental way, and the quality of the process models is quantified by the results of conformance analysis. State-of-the-art conformance analysis techniques are typically optimized and devised for one-time use. However, in process discovery settings which are incremental in nature, it is imperative to have fast conformance analysis. Moreover, the activity logs used for conformance analysis at each stage remain the same. In this paper, we propose an approach that exploits this fact in order to expedite conformance analysis by approximating the conformance results. We use an abstracted version of an activity log, which can be used to compare with the changing (or new) process models in an incremental process discovery setting. Our results show that the proposed technique is able to outperform traditional conformance techniques in terms of performance by approximating conformance scores.

    Originele taal-2Engels
    TitelProceedings - 2018 IEEE 22nd International Enterprise Distributed Object Computing Conference, EDOC 2018
    Plaats van productiePiscataway
    UitgeverijInstitute of Electrical and Electronics Engineers
    Pagina's135-144
    Aantal pagina's10
    ISBN van elektronische versie9781538641392
    DOI's
    StatusGepubliceerd - okt 2018
    Evenement22nd IEEE International Enterprise Distributed Object Computing Conference, EDOC 2018 - Stockholm, Zweden
    Duur: 16 okt 201819 okt 2018

    Congres

    Congres22nd IEEE International Enterprise Distributed Object Computing Conference, EDOC 2018
    Verkorte titelIEEE EDOC 2018
    LandZweden
    StadStockholm
    Periode16/10/1819/10/18

    Vingerafdruk Duik in de onderzoeksthema's van 'Fast conformance analysis based on activity log abstraction'. Samen vormen ze een unieke vingerafdruk.

    Citeer dit