Detecting Novel Behavior and Process Enhancement with Multimodal Process Mining

As a result of industrial developments and revolutions, the importance of data has increased. The reflection of this increasing importance in process mining is the importance of scope expansion with perspective data containing other data attributes, unlike the control perspective. With current studies, analysis is carried out from a multi-perspective with the event logs created by the systems and the recommendation is given regarding the real system. In this study, the production of event logs can be realized with the help of digital twins, another shining area as a result of industrial revolutions. Event log production can be achieved with the process representation modeled without any system. We generate records by passing a process representation using the fault repair process event logs. Additionally, we do not experience the loss of context experienced with multi-perspective analysis, even though we use other perspective data with multi-modal analysis. Multimodal models are created by including other perspective data into the analysis through root cause analysis with the help of metrics, the detection of new behavior, and process enhancement with new models.