A Probabilistic Approach to Event-Case Correlation for Process Mining

Abstract

Process mining aims to understand the actual behavior and performance of business processes from event logs recorded by IT systems. A key requirement is that every event in the log must be associated with a unique case identifier (e.g., the order ID in an order-to-cash process). In reality, however, this case ID may not always be present, especially when logs are acquired from different systems or when such systems have not been explicitly designed to offer process-tracking capabilities. Existing techniques for correlating events have worked with assumptions to make the problem tractable: some assume the generative processes to be acyclic while others require heuristic information or user input. In this paper, we lift these assumptions by presenting a novel technique called EC-SA based on probabilistic optimization. Given as input a sequence of timestamped events (the log without case IDs) and a process model describing the underlying business process, our approach returns an event log in which every event is mapped to a case identifier. The approach minimises the misalignment between the generated log and the input process model, and the variance between activity durations across cases. The experiments conducted on a variety of real-life datasets show the advantages of our approach over the state of the art.

Robotic Process Mining: Vision and Challenges

“…The problem of UI log segmentation is also related to that of correlating uncorrelated events in event logs used for process mining (Bayomie et al 2019(Bayomie et al , 2016Ferreira and Gillblad 2009). However, this problem has been addressed in restrictive settings.…”

“…However, this problem has been addressed in restrictive settings. In particular, Ferreira and Gillblad (2009) addressed the problem when the process (in our case the routine) does not have cycles/repetitions, whereas (Bayomie et al 2016(Bayomie et al , 2019 assume that a process model is given as input, which means that the the routine specification is known. Also, the approaches in Ferreira and Gillblad (2009) and Bayomie et al (2016) were shown to produce rather inaccurate results, whereas RPM seeks to identify routines with high levels of confidence, given that replicating a routine inaccurately can lead to costly errors, especially in contexts where unattended bots are used.…”

“…Indeed, if we want an RPA bot to reproduce the actions of a routine, we need to encode in the bot's script how the parameters of each action are computed from the routine's input parameters or from the parameters of previous actions in the routine. Recent work (Bosco et al 2019) suggests that this step in the discovery of executable routines can be implemented using existing methods for automated discovery of data transformations ''by example'' (Abedjan et al 2016;Jin et al 2017). However, these methods suffer from scalability issues.…”

Exploring the Challenge of Automated Segmentation in Robotic Process Automation

“…Even if more focused on traditional business processes in BPM rather than on RPA routines, Bayomie et al [8] address the problem of correlating uncorrelated event logs in process mining in which they assume the model of the routine is known. Since event logs allow to store traces of one process model only, as a consequence this technique is able to achieve Case 1.1 only.…”

“…In the field of process discovery, Mȃruşter et al [27] propose an empirical method for inducing rule sets from event logs containing execution of one process only. Therefore, as in [8], this method is able to achieve Case 1.1 only, thus making the technique ineffective in presence of interleaved and shared user actions. A more robust approach, developed by Fazzinga et al [12], employs predefined behavioural models to establish which process activities belong to which process model.…”