Applying Process Mining in Manufacturing and Logistic for Large Transaction Data

Process mining is a promising approach to extract actual business processes form event logs. However, process mining algorithms often result in unstructured and unclear process models. Moreover, sufficient data quality is required for accurate interpretation. Therefore, adopting process mining for the field of manufacturing and logistics should take into account the complexity and dynamics as well as the heterogeneous data sources and the quality of event data. Therefore, the objective of this work is to study the application of process mining in the manufacturing and logistics domain with real data from manufacturing companies. We propose a methodology to improve the limitations of process mining by using a Markov chain as a sequence clustering technique in the data preprocessing step and apply heuristic mining to extract the business process models. Finally, we provide results from an experiment with real-world data in which we successfully improve the quality of discovered process model in the regards of replay fitness dimension.