Page 259 -

Chapter 9
Operational Support

Most process-mining techniques work on “post mortem” event data, i.e., they an-
alyze events that belong to cases that have already completed. Obviously, it is not
possible to influence the execution of “post mortem” cases. Moreover, cases that are
still in the pipeline cannot be guided on the basis of “post mortem” event data only.
Today, however, many data sources are updated in (near) real-time and sufficient
computing power is available to analyze events when they occur. Therefore, process
mining should not be restricted to off-line analysis and can also be used for online
operational support. This chapter broadens the scope of process mining to include
online decision support. For example, for a running case the remaining flow time
can be predicted and suitable actions can be recommended to minimize costs.

9.1 Refined Process Mining Framework

Thus far, we identified three main types of process mining: discovery, conformance,
and enhancement (cf. Figs. 1.4 and 8.1). Orthogonal to these types of process mining
we identified several perspectives including: the control-flow perspective (“How?”),
the organizational perspective (“Who?”), and the case/data perspective (“What?”).
The classification of process mining techniques into discovery, conformance, and
enhancement does reflect that analysis can be done online or off-line. Moreover,
Figs. 1.4 and 8.1 do not acknowledge that there are essentially two types of models
(“de jure models” and “de facto models”) and two types of data (“pre mortem” and
“post mortem” event data) [95].
Figure 9.1 shows our refined process mining framework. As before, we assume
some external “world” consisting of business processes, people, organizations, etc.
and supported by some information system. The information system records in-
formation about this “world” in such a way that events logs can be extracted as
described in Chap. 4.
Figure 9.1 emphasizes the systematic, reliable, and trustworthy recording of
events by using the term provenance. This term originates from scientific comput-
W.M.P. van der Aalst, Process Mining, 241
DOI 10.1007/978-3-642-19345-3_9, © Springer-Verlag Berlin Heidelberg 2011