Mining Frequent Instances on Workflows

A workflow is a partial or total automation of a business process,
in which a collection of \emph{activities} must be executed by
humans or machines, according to certain procedural rules. This
paper deals with an aspect of workflows which has not so far
received much attention: providing facilities for the human system
administrator to monitor the actual behavior of the workflow
system in order to predict the ``most probable'' workflow
executions. In this context, we develop a data mining algorithm
for identifying frequent patterns, i.e., the workflow
substructures that have been scheduled more frequently by the
system. Several experiments show that our algorithm outperforms
the standard approaches adapted to mining frequent instances.