Process tree discovery using a probabilistic inductive miner

1. A computer implemented method for splitting an event log into sub-event logs, comprising:

receiving the event log of a process;

calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity;

determining a cut location in the event log based on the activity relation scores; and

splitting the event log into the sub-event logs based on the cut location.

2. The computer implemented method of claim 1 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity.

3. The computer implemented method of claim 2 , wherein:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity comprises dividing the frequency of occurrence of the first activity between occurrences of the second activity by the frequency of occurrence of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity comprises dividing the frequency of occurrence of the second activity between occurrences of the first activity by the frequency of occurrence of the first activity.

4. The computer implemented method of claim 1 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

generating a directly follows graph of the event log and an indirectly follows graph of the event log;

filtering the directly follows graph and the indirectly follows graph; and

calculating the activity relation scores based on the filtered directly follows graph and the filtered indirectly follows graph.

5. The computer implemented method of claim 1 , further comprising:

adding one or more nodes to a process tree for each of the sub-event logs to generate the process tree of the process.

6. The computer implemented method of claim 5 , further comprising:

generating a process model based on the process tree.

7. The computer implemented method of claim 1 , wherein the process is a robotic process automation process.

8. An apparatus comprising:

a memory storing computer instructions for splitting an event log into sub-event logs; and

at least one processor configured to execute the computer instructions, the computer instructions configured to cause the at least one processor to perform operations of:

receiving the event log of a process;

calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity;

determining a cut location in the event log based on the activity relation scores; and

splitting the event log into the sub-event logs based on the cut location.

9. The apparatus of claim 8 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity.

10. The apparatus of claim 9 , wherein:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity comprises dividing the frequency of occurrence of the first activity between occurrences of the second activity by the frequency of occurrence of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity comprises dividing the frequency of occurrence of the second activity between occurrences of the first activity by the frequency of occurrence of the first activity.

11. The apparatus of claim 8 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

generating a directly follows graph of the event log and an indirectly follows graph of the event log;

filtering the directly follows graph and the indirectly follows graph; and

calculating the activity relation scores based on the filtered directly follows graph and the filtered indirectly follows graph.

12. The apparatus of claim 8 , the operations further comprising:

adding one or more nodes to a process tree for each of the sub-event logs to generate the process tree of the process.

13. The apparatus of claim 12 , the operations further comprising:

generating a process model based on the process tree.

14. The apparatus of claim 8 , wherein the process is a robotic process automation process.

15. A computer program embodied on a non-transitory computer-readable medium for splitting an event log into sub-event logs, the computer program configured to cause at least one processor to perform operations comprising:

receiving the event log of a process;

calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity;

determining a cut location in the event log based on the activity relation scores; and

splitting the event log into the sub-event logs based on the cut location.

16. The computer program of claim 15 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity.

17. The computer program of claim 16 , wherein:

comparing a frequency of occurrence of the second activity with the frequency of occurrence of the first activity between occurrences of the second activity comprises dividing the frequency of occurrence of the first activity between occurrences of the second activity by the frequency of occurrence of the second activity; and

comparing a frequency of occurrence of the first activity with the frequency of occurrence of the second activity between occurrences of the first activity comprises dividing the frequency of occurrence of the second activity between occurrences of the first activity by the frequency of occurrence of the first activity.

18. The computer program of claim 15 , wherein calculating an activity relation score for a parallel relationship operator for each respective pair of activities of a plurality of pairs of activities in the event log based on 1) a frequency of occurrence of a first activity of the respective pair of activities between occurrences of a second activity of the respective pair of activities and 2) a frequency of occurrence of the second activity between occurrences of the first activity comprises:

generating a directly follows graph of the event log and an indirectly follows graph of the event log;

filtering the directly follows graph and the indirectly follows graph; and

calculating the activity relation scores based on the filtered directly follows graph and the filtered indirectly follows graph.

19. The computer program of claim 15 , the operations further comprising:

adding one or more nodes to a process tree for each of the sub-event logs to generate the process tree of the process.

20. The computer program of claim 15 , wherein the process is a robotic process automation process.