Data Requirements for Process Learning

Process flexibility and adaptability is essential in environments where the processes are prompt to changes and variations. Process learning is a possible approach for automatically discovering from process log data those process paths that yielded good outcomes and suggesting appropriate process model modifications to enhance future process performance in such environments. We discuss and establish the data requirements for process learning, applicable to clinical process management. Our discussion extends a previously established learning process model (LPM) by providing a formal set of data requirements which enables us to accomplish effective learning. Learning data requirements are illustrated by walking through the application of the LPM framework to a clinical process.