Comparative Process Mining in Nursing-Care: A Case Study using Concurrency Profile Clustering
BSc Thesis Project
Title: Comparative Process Mining in Nursing-Care: A Case Study using Concurrency Profile Clustering
Author: Niklas Dohmen
1st examiner: Prof.Dr.ir Wil M.P. van der Aalst
2nd examiner: Prof.Dr. Markus Strohmaier
Process mining is a relatively new field of data science that tries to extract meaningful insights from data stored during the execution of processes. An important and challenging topic is the discovery of complex spaghetti-like processes that require more advanced techniques due to the unstructured data. Multiple efforts are made in order to derive valuable information out of such complex and often human-made process workflows. In this thesis, we propose an approach to subdivide an event log using time-based concurrency information about events, i.e., we are creating so called concurrency profiles, that are describing the level of multi-tasking inside a process, to cluster the event log accordingly. This helps to analyse process workflows with respect to the level of concurrency and forms the event log in more homogeneous groups in which a similar concurrency workflow took place. To validate our method, we evaluate our approach on a synthetical and two real-life datasets. The latter includes a comparative case study from nursing care of a Dutch hospital. Our results show a weak link between concurrency profiles and process models.