Elusive Case


  • Events in a log are not explicitly linked to their case identifiers
    • Concept of a ‘case’ cannot be trivially defined by using the information in the log as-is
  • Nevertheless, the notion of a ‘case’ exists and can be discerned (by domain experts)


  • Process mining requires each event to be attributable to a case
  • The existence of this imperfection pattern will prevent a process mining analysis being conducted as the concept of a case cannot be applied

Data Quality Issues

I16 - Missing data: relationships
  • The relationship between events and cases is missing

Manifestation and Detection

  • Pattern signature — there dies not exist
    1. a set of attributes (one or more)
    2. common to all events
    3. for which the value of the attribute/s in each event can be used to group events into a case i.e. all events relate to the same process instance and collectively capture all activities of a case
    For the vehicle/GPS tracking data in the table it is not possible (without additional data) to construct cases as journeys, by a vehicle, to one or more areas.


  • Need to correctly relate events with cases, often this can be done through correlation with another source
  • Example: use of journey planner table to identify start and end locations of journeys

Side-effects of Remedy

  • Side effects of such a remedy include (i) incorrect mapping of events to their cases, and (ii) the omission of many events in the log as they cannot be mapped to any particular cases resulting in the loss of potentially important information