Implementation and Evaluation of an Automatic Personal Workflow Extraction Method

  • Authors:
  • Akira Abeta;Ken'ichi Kakizaki

  • Affiliations:
  • -;-

  • Venue:
  • COMPSAC '99 23rd International Computer Software and Applications Conference
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

We report the implementation and evaluation of a method to automatically extract a workflow during a user's ordinary use of a personal information management system (PIM). Our method accumulates the target event and time of the user's operations such as registrations and references. In order to organize related work events, our method extracts relations between events by calculating frequency of a user's sequential reference using a balloon-help-function and connects related work events based on the relations. Furthermore, we improve the accuracy of our organizing method by introducing a function that offers the user the list of the events related to his target event in order to encourage the user to make sequential references to related work events. We have constructed a prototype PIM and tested it with 6 users. We confirmed that our method could make groups of related work events and display them as connected graphs with relations shown. We discuss the efficiency of sharing this extracted workflow using sample graphs.