A Bpmn Data Model to Keep A Multı-Perspectıve Process Model on The Blockchaın
Date
2021Author
Ekici, Burakcan
xmlui.dri2xhtml.METS-1.0.item-emb
6 ayxmlui.mirage2.itemSummaryView.MetaData
Show full item recordAbstract
Among the various data mining techniques, Process Mining is one of the techniques dedicated to analyzing the operational processes based on event logs to improve them. By discovering the process and checking the conformance, Process Mining focuses to enhance the existed process. At that point, the Multi-Perspective Process Mining extends the information, which could be taken from the existed process model, by evaluating the several attributes that have been collected with three essential attributes (i.e., case id, activity, and timestamp). For each perspective, there is a different technique, which is dedicated to the relevant perspective, applied and these techniques may need to consume the results of one another in sequence. Therefore, keeping what is produced from each perspective in a secure and immutable way while applying the multiple perspectives becomes important. In this study, a BPMN-extended Data Model is proposed to hold the models from the Multi-Perspective Process Mining and a tool is developed to keep this Data Model as an asset into a private blockchain developed by using Hyperledger Fabric. Thus, thanks to the advantages presented by blockchain technology (immutability, decentralized, transparency, etc.), this Data Model can be kept invariably, regardless if the environment is trusted. Also, due to its distributed ledger technology and decentralized structure, blockchain guarantees that the system will never be interrupted in any case and governed by anyone. We tested our proposed Data Model on two different datasets that contain different numbers of events, data attributes, and originators, and then we compared the results with what they should have been, with respect to the outputs of ProM tool. For both cases, the Data Model held exactly the same data with what it should have produced in the perspective tested. The tool developed in this study has been containerized by using Docker as ready to use. The tool is available at; https://github.com/burakcanekici/multi-perspective-pm-chain