A Graph-based Method for Detecting Wrong Decision and Wrong Throughput Time Based on BPMN Process Model of Procurement

The documented Standard Operating Procedure (SOP) constitutes the first step in managing organizational business processes. SOP models constructed with the Business Process Model and Notation (BPMN) facilitate a structured representation of business processes. Multiple prior studies have established a graph-based anomaly detection method. Most of these approaches depend on manually constructing process models before converting them into graphs. In numerous instances, the transformation of BPMN models into graph process models is not executed automatically, potentially resulting in inconsistencies that diminish the precision of anomaly detection. This study proposes detecting wrong decisions and detecting wrong throughput time, concentrating on tasks frequently not explicitly identifiable in traditional event logs a graph-based method to automate the conversion of BPMN business process models into graph process models. This proposed approach has been evaluated using a dataset of 56 cases, and it achieved an accuracy rate of 100% in identifying wrong decision and wrong throughput time. The experimental results achieved high accuracy; however, further validation using heterogeneous datasets is required to confirm generalization.