Title: Event Log Imperfection Patterns for Process Mining
Presented by Arthur HM ter Hofstede
Based on joint work with Suriadi Suriadi, Rob Andrews and Moe Wynn
An essential component of a process mining exercise is the ability
to properly prepare the log(s) for analysis. This entails detecting various
types of errors and deficiencies as the quality of the analysis heavily
depends on the quality of the event log. Unfortunately this is not a trivial task
and much depends on the skills and experiences of the analyst(s) involved.
In order to more systematically approach log preparation and to
reduce the dependency on specialist skills, a set of patterns is proposed.
These patterns aim to capture the various types of issues that may be encountered,
the way they may manifest themselves, how they may influence analysis results,
and potential remedies to counter these issues.