Holistic discovery of decision models from process execution data

Abstract

Abstract The analysis of business processes is a multifaceted problem that is comprised of analysing both activities’ workflow, as well as the decisions that are made throughout that workflow. In process mining, the automated discovery of process models from event data, a strong emphasis can be found towards discovering this workflow, as well as how data influences that workflow, i.e., decision point analysis. Nonetheless, the data that is pertaining to the activities in the workflow does not necessarily correlate with the control flow. Decisions that influence variables that are used by activities can also impact other variables used later in the workflow without interfering with the order in which activities are executed. Discovering this has not been addressed in literature, as current decision mining techniques still rely on control flow. To address this, Process Mining Integrating Decisions (P-MInD) is proposed. It relies on uncovering the influence of activities on their variables and connects them by making use of sequence dependencies present in the data. Furthermore, it allows to find autocorrelations, as well to incorporate case variables. This allows to establish a holistic image of the decision layer, captured with Decision Model and Notation (DMN), that is consistent with the discovered control flow. Furthermore, decision model conformance checking, i.e., the matching of event logs with holistic models, is proposed to offer a way to verify whether the models are corresponding with the behaviour that is present in the current system. P-MInD is implemented and used on real-life data to verify its effectiveness.