A Framework for Estimating Simplicity of Automatically Discovered Process Models Based on Structural and Behavioral Characteristics

Abstract

A plethora of algorithms for automatically discovering process models from event logs has emerged. The discovered models are used for analysis and come with a graphical flowchart-like representation that supports their comprehension by analysts. According to the Occam's Razor principle, a model should encode the process behavior with as few constructs as possible, that is, it should not be overcomplicated without necessity. The simpler the graphical representation, the easier the described behavior can be understood by a stakeholder. Conversely, and intuitively, a complex representation should be harder to understand. Although various conformance checking techniques that relate the behavior of discovered models to the behavior recorded in event logs have been proposed, there are no methods for evaluating whether this behavior is represented in the simplest possible way. Existing techniques for measuring the simplicity of discovered models focus on their structural characteristics such as size or density, and ignore the behavior these models encoded. In this paper, we present a conceptual framework that can be instantiated into a concrete approach for estimating the simplicity of a model, considering the behavior the model describes, thus allowing a more holistic analysis. The reported evaluation over real-life event logs for several instantiations of the framework demonstrates its feasibility in practice.

Responsible Process Mining

“…The issue with initially proposed quality measures led to several new methods and definitions for measuring various model quality dimensions being proposed [30][31][32][33][34]. Main complications for model quality in process mining are that process models commonly exhibit infinite behaviour (through loops) and the absence of negative examples, i.e., behaviour that the model should not contain [1].…”

Stochastic Process Discovery by Weight Estimation

“…(3) Petri net entity count (places and transitions) and ( 4) edge count are used as structural simplicity measures, ensuring that conformance quality has not been achieved by sacrificing model simplicity and comprehensibility. Entity and arc counts have existing uses in process model evaluation [14,17], and were preferred here over behavioural simplicity measures [16], though these measures also have limitations, including specificity to Petri nets, and insensitivity to the stochastic perspective of GSPNs. The duration of a discovery process was also captured, and direct discovery times are compared with combined runtimes for discovery and estimation.…”

“…Unlike enhancement techniques, estimators can potentially change control flows when producing a stochastic process model. Stochastic process models have a corresponding, emerging, set of stochastic process conformance measures [16,20,21]. Consequently, the algorithms and models presented here are evaluated, in Sect.…”

Tuning Machine Learning to Address Process Mining Requirements

“…The granularity gap between these events and the activities considered by classic PM analysis has often been bridged using ML models [8,9] that compute virtual activity logs, a problem which is also known as log lifting [10]. ML has been proposed as a key technology to strengthen existing techniques, for example, using trace clustering to reduce the diversity that a process discovery algorithm must handle in analyzing an event log [11,12,13,14], to simplify the discovered models [15,16,17], or to support real-time analysis on FIGURE 1. The PM tasks and their relation to ML event streams [18,19,20].…”