Evaluating Process Discovery From Loop Structures

Abstract

Modern organizations increasingly rely on sophisticated information systems to manage their business processes, generating detailed event logs that record key activities. Process mining, and specifically process discovery, utilizes these event logs to construct models that represent the underlying processes. Effective process discovery is crucial for organizations to gain insights into their operations, identify inefficiencies, and drive continuous improvement. This study evaluates the capability of four process discovery algorithms – Alpha Miner, Heuristics Miner, ILP Miner, and Inductive Miner – in handling complex workflow patterns, particularly those involving intricate loop and nested control-flow structures. By generating synthetic event logs and applying the PM4Py library, we assess the algorithms' performance using metrics such as fitness, precision, generalization, and simplicity. Our results highlight the strengths and limitations of each algorithm, providing valuable insights for researchers and practitioners in the field.