Statistical Sequence Analysis for Business Process Mining and Organizational Routines
Breuker Dominic, Matzner Martin
Analyzing discrete event sequences has become a popular field in recent years. In the area of business process mining, numerous techniques have been developed to discover the structure of business processes by means of traces they leave behind in information systems. In organizational routines literature, these traces have been identified as a valuable source of information to investigate the dynamics of routines and how they evolve over time. However, both areas have been discussed in separation only. But in both areas alike the fundamental problem is to acquire knowledge about regularities in sequences of events based on observations thereof, and thus, we argue that process mining has the potential to advance research on organizational routines. As with any data analysis problem, one has to deal with problems due to noisy data and small samples. Thus, we show in this paper how to apply simple statistical tools to pattern detection in sequences. Subsequently, we integrate this into the popular algorithm. This paves the way for statistically controlling the risk of falling for erroneous results. To the best of our knowledge, no process mining algorithm is capable of doing this. We are convinced that this will facilitate applicability in organizational routines studies.
Business Process Mining, Organizational Routines, Organizational Genetics, Sequence Analysis