The Web Science Trust

Evaluating implicit judgements from Image search interactions

Smith, Gavin and Ashman, Helen (2009) Evaluating implicit judgements from Image search interactions. In: Proceedings of the WebSci'09: Society On-Line, 18-20 March 2009, Athens, Greece.

[img] PDF (preprint) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
68Kb
[img]
Preview
PDF (Final Version) - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
79Kb

Abstract

The by-product of search engine interactions - the records of what users have clicked as a result of a query - known as click through data'' is an increasingly popular resource of implicit feedback. Recently, however, the validity of such feedback has been questioned. However, the criticisms have focused almost solely on traditional web search, for which users are only presented with short abstracts to evaluate the resource before clicking - and creating the implicit feedback'' between the query and resource. In contrast other search types vary significantly in what the user is presented with and hence what is used to pass judgment and create the implicit feedback''. Of note is web-based image search interactions where the user is presented with a reduced thumbnail of the whole image. Image search click-through data has also been shown to have many potential uses in recent works, however, unlike traditional web search interactions there has been no previous research that has explicitly looked into its validity as implicit judgments. This work addresses this issue, comparing it to what has been discovered for traditional web search. While conceptually similar, the results of previous studies based on web search interactions can not be directly transferred to image search interactions due to the aforementioned difference in the summary displayed and used by the user in the decision to click.

Item Type:Conference or Workshop Item (Paper)
Uncontrolled Keywords:Implicit judgment, Click-though data, Search
Subjects:Web Science Events > Web Science 2009
ID Code:148
Deposited By: W S T Administrator
Deposited On:24 Jan 2009 08:45
Last Modified:25 Oct 2011 16:32

Repository Staff Only: item control page

EPrints Logo
Web Science Repository is powered by EPrints 3 which is developed by the School of Electronics and Computer Science at the University of Southampton. More information and software credits.