search

pm4py / pm4py-core

Public repository for the PM4Py (Process Mining for Python) project.
https://pm4py.fit.fraunhofer.de
GNU General Public License v3.0
722 stars 286 forks source link

Possible issue with IMDFb #82

Closed Javert899 closed 5 years ago

Javert899 commented 5 years ago

A potential problem with IMDFb was signaled on this log <a,b,c> <a,c,b> <c,a,b>

Checks will be done

Javert899 commented 5 years ago

The problem is recognised and actually resolved in the "hotfixes" branch. The expected behavior on the log is to find a parallel cut between ->(a,b) and c.

The cut was correctly recognised by the algorithm, but was discarded due to an additional check on short loops. In some conditions, this is useful to avoid detecting almost-parallel-cuts that shows no connection with the initial node (of the parallel cut). This condition happens, for example, in the a32f0n00 log between the "g" and the "h9" activities, where "g" and "h9" would be detected as a parallel cut satisfying all the other conditions, but actually are in a short loop.

The check that was implemented in PM4Py on short loops was failing when the parallel cut actually starts the tree.

The fix will be released in the next release of PM4Py.