Discovering Data-Aware Declarative Process Models from Event Logs

Abstract

Abstract. A wealth of techniques are available to automatically discover business process models from event logs. However, the bulk of these techniques yield procedural process models that may be useful for detailed analysis, but do not necessarily provide a comprehensible picture of the process. Additionally, barring few exceptions, these techniques do not take into account data attributes associated to events in the log, which can otherwise provide valuable insights into the rules that govern the process. This paper contributes to filling these gaps by proposing a technique to automatically discover declarative process models that incorporate both control-flow dependencies and data conditions. The discovered models are conjunctions of first-order temporal logic expressions with an associated graphical representation (Declare notation). Importantly, the proposed technique discovers underspecified models capturing recurrent rules relating pairs of activities, as opposed to full specifications of process behavior -thus providing a summarized view of key rules governing the process. The proposed technique is validated on a real-life log of a cancer treatment process.

Process mining in oncology: A literature review

“…The methodology followed by the other ten papers [3], [11], [20], [29], [34], [37], [40], [44], [61], [62] was using available plug-ins and/ or functionalities in existing tools to solve the problem in their case study.…”

“…Some papers identified data improvements [28]- [30], [36]- [38], [48], [57], [64] in terms of improving data quality, dimensionality, and complexity. A larger number of important technical improvements were identified in 20 papers [3], [21]- [23], [29]- [31], [35], [39]- [41], [44], [48]- [52], [58], [60], [64]. These were to:…”

Discovering Target-Branched Declare Constraints

“…Other works [10,3,5,7,17,15,14] focus on the discovery of Declare models. The algorithms proposed in [5,17,15] are suitable for discovering standard Declare…”

“…However, these approaches can be hardly used in real-world settings since they are based on supervised learning techniques requiring negative examples. In the work proposed in [14], a first-order variant of LTL is used to specify a limited version of data-aware patterns. Such extended patterns are used as the target language for a process discovery algorithm, which produces data-aware Declare constraints from raw event logs.…”

Performance Analysis of Business Processes using Process Mining

“…The add ref. [26] initial coated the info perspective in declarative process mining, though this approach solely permits for the discovery of discriminative activation conditions. In essence, the main focus of the said approaches is control-flow with extensions to hide knowledge while not analysing resource-related data.…”