Page 263 -

9.2 Online Process Mining 245

Note that the detect and check activities are similar except for the event data used.
The former activity uses “pre mortem” data and aims at online analysis to be able
to react immediately when a discrepancy is detected. The latter activity uses “post
mortem” data and is done off-line.

9.1.3 Navigation

The last category of process mining activities aim at business process navigation.
Unlike the cartography and auditing activities, navigation activities are forward-
looking. For example, process mining techniques can be used to make predictions
about the future of a particular case and guide the user in selecting suitable actions.
When comparing this with a car navigation system from TomTom or Garmin, this
corresponds to functionalities such predicting the arrival time and guiding the driver
using spoken instructions. In Chap. 13, we elaborate on the similarities between car
navigation and process mining.
Figure 9.1 lists three navigation activities: explore, predict, and recommend.

• Explore. The combination of event data and models can be used to explore busi-
ness processes at run-time. Running cases can be visualized and compared with
similar cases that were handled earlier.
• Predict. By combining information about running cases with models (discovered
or hand-made), it is possible to make predictions about the future, e.g., the re-
maining flow time and the probability of success.
• Recommend. The information used for predicting the future can also be used to
recommend suitable actions (e.g., to minimize costs or time). The goal is to enable
functionality similar to the guidance given by car navigation systems.
In earlier chapters, we focused on activities using historic (“post mortem”) data
only, i.e., activities discover, enhance, and check in Fig. 9.1. In the remainder of this
chapter, we shift our attention to online analysis also using “pre mortem” data.

9.2 Online Process Mining

Traditionally, process mining has been used in an off-line fashion using only “post
mortem” data. This means that only completed cases are being considered, i.e., the
traces in the event log are complete traces corresponding to cases that were fully
handled in the past. For operational support we also consider “pre mortem” event
data and respond to such data in an online fashion. Now only running cases are
considered as these can, potentially, still be influenced. A running case may still
generate events. Therefore, it is described by a partial trace.
Figure 9.2 shows the essence of operational support. Consider a case for which
activities a and b have been executed. Partial trace σ p = a,b describes the known
past of the case. Note that the two events may have all kinds of attributes (e.g.,