Data-Driven Decision Support for Business Processes: Causal Reasoning and Discovery

Various types of decisions influence the execution of a business process, e.g., in terms of control-flow and resource assignments. Data recorded during process execution can be used to identify which decisions are informed by data and by previous decisions, to predict their outcome, and to guide interventions as part of a what-if analysis. The latter requires causal models that explain the factors that influence decisions. Yet, existing causal techniques for business processes are limited: they focus on control-flow decisions only, ignore data variables, and use ad-hoc methods to resolve causal conflicts. In this paper, we fill this gap by introducing a causal decision modeling framework that incorporates variables, which allows us to uncover, for example, confounding effects, and capture resource decisions. Moreover, we provide a process-aware causal discovery algorithm, based on the notion of temporal tiers, that takes process precedence into account, without the need for heuristic conflict resolution between process discovery and causal discovery. We demonstrate the effectiveness of our approach through experiments using synthetically generated data, and show a proof-of-concept implementation on a real-world dataset.