Automaatne äriprotsesside avastamine: kirjanduse ülevaade ning võrdlev hindamine koostöös domeeniekspertidega
Abstract
Äriprotsesside kaeve meetodi võimaldavad analüütikul kasutada logisid saamaks teadmisi protsessi tegeliku toimise kohta. Neist meetodist üks enim uuritud on automaatne äriprotsesside avastamine. Sündmuste logi võetakse kui sisend automaatse äriprotsesside avastamise meetodi poolt ning väljundina toodetakse äriprotsessi mudel, mis kujutab logis talletatud sündmuste kontrollvoogu. Viimase kahe kümnendi jooksul on väljapakutud mitmeidki automaatseid äriprotsessi avastamise meetodeid balansseerides erinevalt toodetavate mudelite skaleeruvuse, täpsuse ning keerukuse vahel. Siiani on automaatsed äriprotsesside avastamise meetodid testitud ad-hoc kombel, kus erinevad autorid kasutavad erinevaid andmestike, seadistusi, hindamismeetrikuid ning alustõdesid, mis viib tihti võrdlematute tulemusteni ning mõnikord ka mittetaastoodetavate tulemusteni suletud andmestike kasutamise tõttu. Eelpool toodu mõistes sooritatakse antud magistritöö raames süstemaatiline kirjanduse ülevaade automaatsete äriprotsesside avastamise meetoditest ja ka süstemaatiline hindav võrdlus üle nelja kvaliteedimeetriku olemasolevate automaatsete äriprotsesside avastamise meetodite kohta koostöös domeeniekspertidega ning kasutades reaalset logi rahvusvahelisest tarkvara firmast. Kirjanduse ülevaate ning hindamise tulemused tõstavad esile puudujääke ning seni uurimata kompromisse mudelite loomiseks nelja kvaliteedimeetriku kontekstis. Antud magistritöö tulemused võimaldavad teaduritel parandada puudujäägid meetodites. Samuti vastatakse küsimusele automaatsete äriprotsesside avastamise meetodite kasutamise kohta väljaspool akadeemilist maailma.
Process mining methods allow analysts to use logs of historical executions of business processes in order to gain knowledge about the actual performance of these processes.One of the most widely studied process mining operations is automated process discovery.An event log is taken as input by an automated process discovery method and produces a business process model as output that captures the control-flow relations between tasks that are described by the event log.Several automated process discovery methods have been proposed in the past two decades, striking different tradeoffs between scalability, accuracy and complexity of the resulting models.So far, automated process discovery methods have been evaluated in an ad hoc manner, with different authors employing different datasets, experimental setups, evaluation measures and baselines, often leading to incomparable conclusions and sometimes unreproducible results due to the use of non-publicly available datasets.In this setting, this thesis provides a systematic review of automated process discovery methods and a systematic comparative evaluation of existing implementations of these methods with domain experts by using a real-life event log extracted from a international software engineering company and four quality metrics.The review and evaluation results highlight gaps and unexplored tradeoffs in the field in the context of four business process model quality metrics.The results of this master thesis allows researchers to improve the lacks in the automated process discovery methods and also answers question about the usability of process discovery techniques in industry.
Process mining methods allow analysts to use logs of historical executions of business processes in order to gain knowledge about the actual performance of these processes.One of the most widely studied process mining operations is automated process discovery.An event log is taken as input by an automated process discovery method and produces a business process model as output that captures the control-flow relations between tasks that are described by the event log.Several automated process discovery methods have been proposed in the past two decades, striking different tradeoffs between scalability, accuracy and complexity of the resulting models.So far, automated process discovery methods have been evaluated in an ad hoc manner, with different authors employing different datasets, experimental setups, evaluation measures and baselines, often leading to incomparable conclusions and sometimes unreproducible results due to the use of non-publicly available datasets.In this setting, this thesis provides a systematic review of automated process discovery methods and a systematic comparative evaluation of existing implementations of these methods with domain experts by using a real-life event log extracted from a international software engineering company and four quality metrics.The review and evaluation results highlight gaps and unexplored tradeoffs in the field in the context of four business process model quality metrics.The results of this master thesis allows researchers to improve the lacks in the automated process discovery methods and also answers question about the usability of process discovery techniques in industry.