Evaluating implicit measures to improve web searchACM Trans. Inf. Syst., Vol. 23, No. 2. (April 2005), pp. 147-168.
|
Reviews
[Write a review of this article]
There are no reviews of this article
Notes for this articleModels user satisfaction with a query or with an entire search trail using features extracted by instrumenting the browser.
Potential useful implicit feedback includes: examination (looking at a page), retention (bookmarking a page), and reference from [Oaard and Kim 1998]. Features included dwelling time, scrolls, position of a click in the result set, exit type, image count in target page, etc.
Two important features are dwelling time and exit-type. Printing and bookmarking are correlated with satisfaction too but they are very rare.
Also uses "gene analysis" (similar to the general patterns of [Catledge and Pitkow 1995]). These are general patterns such as query-resultspage-click-resultspage-click-exit or query-resultspage-click-exit. A single query-resultspage-click-exit is a good predictor of satisfaction.
Find related articles from these CiteULike users
Find related articles with these CiteULike tags
BibTeX record
RIS record