Li, Guangming, de Murillas, Eduardo González López, de Carvalho, Renata Medeiros & van der Aalst, Wil M.P. (2018). Extracting object-centric event logs to support process mining on databases. In H. Mouratidis & J. Mendling (Eds.), Information Systems in the Big Data Era – CAiSE Forum 2018, Proceedings (pp. 182-199). (Lecture Notes in Business Information Processing, No. 317). Springer.
Process mining helps organizations to investigate how their operational processes are executed and how these can be improved. Process mining requires event logs extracted from information systems supporting these processes. The eXtensible Event Stream (XES) format is the current standard which requires a case notion to correlate events. However, it has problems to deal with object-centric data (e.g., database tables) due to the existence of one-to-many and many-to-many relations. In this paper, we propose an approach to extract, transform and store object-centric data, resulting in eXtensible Object-Centric (XOC) event logs. The XOC format does not require a case notion to avoid flattening multi-dimensional data. Besides, based on so-called object models which represent the states of a database, a XOC log can reveal the evolution of the database along with corresponding events. Dealing with object-centric data enables new process mining techniques that are able to capture the real processes much better.