Homonymous Label

Description

  • This pattern describes a situation where an activity is repeated multiple times within a case (same activity label applied to each occurrence of the activity), but the interpretation of the activity, from a process perspective, differs between first and subsequent occurrences

Affect

  • The existence of this pattern will result in discovered process models painting an incomplete picture of the process being analysed. Incomplete because the repeated activities, which actually have different meanings, are ‘grouped’ into one, thus, ‘hiding’ certain process information. This is because most, if not all, process discovery algorithms do not treat repeated activities as separate activities in the discovered model.

Data Quality Issues

I2 - Imprecise data: activity name
  • The activity names are too coarse to reflect the different connotations associated with the recorded events

Manifestation and Detection

  • ‘Triage Assessment’ activity occurs multiple times in the case
  • 2nd and 3rd occurrences after Discharged event hence triage is unlikely
  • These occurrences most likely represent a medical officer reviewing the triage assessment results one week after the patient was discharged
  • Pattern signature — (i) existence of an activity within a discovered process model that has many incoming arcs, often including a self-loop arc and arcs from other activities.

    Each repetition of the ‘Correspondence’ activity may refer to correspondence of a different type (e.g. sending a quote, scheduling a meeting, etc).
    (ii) A second indicator is the ratio of the number of times a particular activity occurs in a log and the total number of cases in the log. A high ratio indicates that the activity is repeated many times within a case and requires investigation to check if different interpretations are possible for each repetition of the activity.

Remedy

  • Re–label repeated activities with context sensitive name (e.g. ‘Triage Assessment (Review)’)
  • To do so first identify context when activities are repeated, then develop an approach to re–labeling by adding context information

Side-effects of Remedy

  • May lead to too many distinct activity names thus affecting model readability and making analysis more complex and involved.