Space-Time Cube Operations in Process Mining
Process mining techniques provide data-driven visualizations that help gaining multi-perspective insights into business processes. These techniques build on a variety of algorithms, however without any explicit reference to the spectrum of potential analy
- PDF / 1,835,006 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 4 Downloads / 213 Views
1 Introduction Organizations increasingly use process mining techniques for analyzing their processes based on process execution data (i.e. event-log) [1, 2]. These techniques can be used to discover the process model, check conformance between the data and the process model, explore variants, analyze bottlenecks or predict process performance indicators. The broad spectrum of techniques results from the different inputs (event-log, process model, etc.) and the different outputs that are generated (various types of metrics, charts, etc.). For instance, process discovery techniques use event-logs only to construct a process model that best fits the data, while conformance checking techniques use eventlogs and a process model to project the results onto the model. All process mining techniques have in common that they generate visualizations from process data. These visualizations are beneficial to process analysts for process optimization, monitoring, and decision making. One of the current challenges of process mining is the richness and diversity of techniques. The 6.9 distribution of ProM from May 2020 alone contains 269 packages and most of them provide separate analysis functionality. On the one hand, it is unclear c IFIP International Federation for Information Processing 2020 Published by Springer Nature Switzerland AG 2020. All Rights Reserved J. Grabis and D. Bork (Eds.): PoEM 2020, LNBIP 400, pp. 405–414, 2020. https://doi.org/10.1007/978-3-030-63479-7_28
406
D. Bayomie et al.
which techniques offer comparable insights. On the other hand, it is unclear if opportunities for analysis techniques have been missed. What is absent at this stage is a framework that helps organizing the spectrum of techniques and inspiring the development of new techniques. In this paper, we address this research gap by referring to research on information visualization. More specifically, we build on the space-time cube proposed in [3] as a framework for identifying abstraction operations that produce two-dimensional visualizations from multi-dimensional data. Using this framework, we analyze to which extent commercial process mining tools instantiate those operations. We find that the majority of the operations are already supported by the tools, but there are still unsupported ones, which exhibit opportunities for future research and tool innovation. The remainder of this paper is organized as follows. Section 2 describes the background of process mining techniques and space-time cube operations. Section 3 presents our findings of analyzing the spectrum of analysis techniques offered by two commercial process mining tools. Finally, Sect. 4 concludes our work and provides directions for future research.
2 Background This section discusses the background of process mining and gives an overview of the space-time cube and its operations. 2.1
Process Mining and Process Analytics
Process mining techniques enable companies to discover, monitor, and improve real business processes by extracting information from event-logs of pro
Data Loading...
