Information Technology Reference
In-Depth Information
Step 7: Assemble logs. Illustrate common and alternative process paths as process view instances. Con-
sider thereby the entire process view (from start to the end) as one process instance. If there are
decisions included, then transform the common and alternative process paths into different
instances.
The next task is to prepare the log instances of the entire process including all
process views. The output is a process model of the entire process that captures possi-
ble variations of process paths by considering the allowed decision combinations. For
preparing such logs, the following preconditions need to be considered.
Precondition 1: The process that is analyzed is clearly defined and has a common thread (e.g. in a sales
process, the common thread is the order on which the process participants work on
throughout the whole process).
Precondition 2: The process views are already transformed into process view logs. The process view logs
reflect common and alternative process paths.
The steps necessary to merge process view logs in such way that they reflect the
whole process with its alternative paths are presented in the following. In addition to
the mentioned steps below, exchange with the process participants during workshops
might support the analyst to get an immediate “big picture” of the whole process and
to reduce errors and misinterpretations in the whole process model.
Step 1: Determine the number of process instances (cases). The process instances illustrate possible entire
process path. Elaborate per process instance which decision path of what process member is con-
sidered. Consider rules and conditions when assembling the process view instances.
Step 2: Bring the process fragment scenarios together and sort the operations according to the time
stamps.
In this work our intension is not to present a new mining algorithm for process view
mining. Our goal is rather to discuss the purpose and benefits of mining prepared logs.
Mining of the process view logs and the entire process supports the designer in
constructing and visualizing process models from data analyzed and prepared in a
structured and traceable way. The mining of each single log of the individual process
PROCESS VIEW MERGING
PROCESS VIEW AGGREGATION
Organizational
Information
Merged PVs
(Path1 of the
entire process):
Case 1:
START(A)
END (A)
START (B)
END(B)
START (D)
…
Process Mining:
Process View Mining:
Case 3:
START(A)
END(A)
START(B)
START(C)
END(C)
END(B)
PV Case 3:
START(A)
END(A)
START(B)
START(C)
END(C)
END(B)
To D o 's
Case 2:
START(A)
END(A)
START(B)
START(C)
END(C)
END(B)
PV Case 2:
START(A)
END(A)
START(B)
START(C)
END(C)
END(B)
Person 1:
START(A)
END(A)
START(B)
END(B)
START(D)
END(D)
Person 2:
START(A)
END(A)
START(C)
START(D)
END(C)
END(D)
Person 3:
START(E)
START(F)
START(G)
END(E)
END(F)
END(G)
Case 1:
START(A)
END(A)
START(B)
END(B)
START (D)
…
Process View (PV)
Case 1:
START(A)
END(A)
START(B)
START(C)
END(C)
END(B)
…
Event Logs
Event Logs
View Person 1
View Person 2
View Person 3
A
A
A
Merged Process
Model:
B
D
B
C
D
E
F
G
Aggregated
Process Model:
B
D
A
D
E
E
B
C
A
D
E
Process (View) Models
Merged Process
(View) Models
C
Fig. 2.
Process view aggregation vs. merging
