Inferring Editor Roles in Ontology Engineering Projects - A Large-Scale Study of WebProtégé Change Logs

Tracking #: 2015-3228

This paper is currently under review
Authors: 
Patrick Kasper
Simon Walk
Matthew Horridge
Denis Helic
Mark A. Musen

Responsible editor: 
Marta Sabou

T
Submission type: 
Full Paper
Abstract: 
Previous research attempting to analyze and understand the processes involved in ontology creation was often limited in scope and generality due to the lack of available data. In this paper, we shed light on the editing behavior of users creating ontologies by investigating change logs for a large number of ontology engineering projects. To that end, we analyze a corpus of nearly five hundred ontology engineering project change logs, extracted from the Web-based online ontology editing tool WebProtégé. The change logs contain over four million edits made by over one thousand users. In our analysis, we cluster users with similar editing behavior by applying k-means clustering on their editing sequences. We infer and describe five distinct editor roles, revealing that individual users concentrate on specific tasks for extended periods. We further investigate these individual clusters by (i) analyzing their distributions over the projects, and by (ii) tracking editor role changes over the complete lifespan of the individual projects. Our results indicate that the majority of projects have one leading editor role and that there are regular patterns of how users switch between roles during different phases of an ontology engineering project. Moreover, our results reveal valuable insights into the engineering processes and the editor role distribution and evolution of nearly five hundred real-world ontology engineering projects, which can potentially be leveraged for improving existing ontology editing tools by, for example, creating automatically adapting interfaces to support the individual editor roles.
Full PDF Version: 
Tags: 
Under Review