Inductive miner com agrupamento de sublogs de nós fall-through

Abstract

The increasing complexity of business processes, coupled with extensive data collection through real-world event logs, has given rise to unstructured process models commonly referred to as “spaghetti” process models. These models, characterized by a lack of clarity and structure, pose a significant challenge in understanding and optimizing the underlying processes. Conventional approaches often result in intricate and difficult-to-interpret representations, negatively impacting operational efficiency and decision-making. In response to this scenario, this work addresses the existing gap in the simplification and understanding of these unstructured models through the algorithm named IM_Cluster. The focus is on clustering trace fragments associated with fall-throughs as an innovative strategy to extract patterns and reveal underlying structures. The obtained results demonstrate that the IM_Cluster proposal yields significant outcomes when compared to some works in the literature, solidifying its effectiveness as an approach for process model simplification. The central problem lies in the need to develop an effective approach that not only deals with the inherent complexity of these models but also provides a clear and cohesive view of the underlying processes, thereby promoting more efficient and informed management.