Behavioural Similarity Measurement of Business Process Model to Compare Process Discovery Algorithms Performance in Dealing with Noisy Event Log

Process discovery algorithms have different strength and weakness to find the most suitable model. The five process discovery algorithms will be compared in this research such as Alpha, Heuristic Miner, Duplicate Genetic, Genetic, and Inductive Miner, to get the recommendation of chosen algorithm in modeling business process from Shoes Manufacturing Company. This research provides two case studies of business process, i.e. the business process of planning-to-stock and production planning-to-export. This research focuses on how the performance and ability of process mining algorithm in facing event logs which consist of 1% noise. This research will rank those five algorithms based on behavioral similarity values between reference and mined model. Result from the behavioral similarity measurement shows that the Genetic and Inductive Miner Algorithm is recommended for planning-to-stock business process, whereas inductive Miner algorithms is recommended for production planning-to-export business process in Shoes Manufacturing Company. (C) 2019 The Authors. Published by Elsevier B.V.