2008 | OriginalPaper | Buchkapitel
Abductive Workflow Mining Using Binary Resolution on Task Successor Rules
verfasst von : Scott Buffett
Erschienen in: Rule Representation, Interchange and Reasoning on the Web
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
The notion of
abductive workflow mining
is introduced, which refers to the process of discovering important workflows from event logs that are believed to cause or explain certain behaviour. The approach is based on the notion of abductive reasoning, where hypotheses are found that, if added to a rule base, would necessarily cause an observation to be true. We focus on the instance of workflow mining where there are critical tasks in the underlying process that, if observed, must be scrutinized more diligently to ensure that they are sufficiently motivated and executed under acceptable circumstances. Abductive workflow mining is then the process of determining activity that would necessarily imply that the critical activity should take place. Whenever critical activity is observed, one can then inspect the abductive workflow to ascertain whether there was sufficient reason for the critical activity to occur. To determine such workflows, we mine recorded log activity for task successor rules, which indicate which tasks succeed other tasks in the underlying process. Binary resolution is then applied to find the abductive explanations for a given activity. Preliminary experiments show that relatively small and concise abductive workflow models can be constructed, in comparison with constructing a complete model for the entire log.