On this webpage, where would you click to find the Career Center?
The semantic content of the various links will likely influence your decision.
Picture derived from the Seton Hill University website in February, 2004.
Ishwinder Kaur's research focuses on exploring how the semantic content of a webpage influences visual search. For example, if a user is looking for information about services offered by the Career Center, will it matter if the actual information is behind the link labelled 'Student life' and not the one labelled 'Centers'? Furthermore, how would the user's search behavior be affected by the alternative links available on the page? Specific research questions include:
A primary first question revolves around defining and quantifying semantic relatedness. We are currently exploring how different tools for assessing semantic relatedness, including Latent Semantic Analysis and WordNet, predict actual search behavior. The goal of the research is to model the cognition and behavior of visual search.
An understanding of how semantic content affects visual search will be useful in designing a predictive tool of user behavior on a webpage. Previous work has focused primarily on the visual appearance and layout of text on a webpage (e.g., font size, font color). If we can predict and simulate human visual search, the tool would be useful in evaluating the usability of a visual interface (e.g., a webpage).
Kaur, I. & Hornof, A. J. (2005). A comparison of LSA, WordNet and PMI-IR for predicting user click behavior. Proceedings of ACM CHI 2005: Conference on Human Factors in Computing Systems, New York: ACM, 51-60. Available as a PDF file.
Last updated: 1/19/2008 by ajh