IRES: A User-Centred Approach to the use of Eyetracking in Information Retrieval Systems – Københavns Universitet

IRES: A User-Centred Approach to the use of Eyetracking in Information Retrieval Systems

Specialeforsvar ved Michael Jensen

ABSTRACT:

One way of improving search results is to examine what people look at and
use this to improve subsequent searches. In my thesis I implemented an
eyetracking system for real-time relevance feedback. The system used fixated
words and term frequency/inverse document frequency for query expansion. I
tested the system using a user-centred approach that included measuring user
satisfaction with the system. I found indicators that the system returned
more relevant results, but that users were less satisfied with the system
than a baseline system. However, none of the results were statistically
significant. I found many challenges in testing with a user-centred approach
and gave advice for further work using this approach with eyetracking
systems for relevance feedback.

Eksaminator:

Kasper Hornbæk

Censor:

Morten Hertzum, RUC