Year of Publication

1999

Paper Type

Master's Thesis

College

College of Computing, Engineering & Construction

Degree Name

Master of Science in Computer and Information Sciences (MS)

Department

Computing

First Advisor

Dr. Krissten N. Cooper

Second Advisor

Dr. Kenneth E. Martin

Third Advisor

Dr. Robert F. Roggio

Abstract

The objective of this research is to determine the effects over time of a dynamic system that adapts itself to a user's current state of expertise, in terms of the application domain, by constantly monitoring the user throughout use of the system, placing them in appropriate user models when this expertise has changed.

A dynamic system, named ER-by-Design version 2.0, is presented, consisting of an inference component, a help system, a help/assistance screen, and user models. The user models are responsible for adapting the system interface to the level of expertise of the user. The system monitors and analyzes a user's interactions in order to evaluate user expertise, placing the user in the most appropriate model based on this evaluation.

Through analysis of data collected from participants' sessions with both versions of the system, it is shown that over time, through the use of ER-by-Design version 2.0, users accessed help less often and perceived the system as more beneficial when compared to a system with a static, generic interface. In addition, users who had the least experience with ER modeling concepts created more correct diagrams with ER-by-Design version 2.0 than with a static version of the system.

Share

COinS