Comparing Tag Clouds, Term Histograms, and Term Lists for Enhancing Personalized Web Search

Abstract

Although static ranked lists remain the dominant Web search interface, they can limit the ability of Web searchers to find desired information when it is buried deep in the collection of search results. Web search visualization and Web search personalization are two active research directions that have shown promise for improving the user experience while searching the Web. In this paper, we propose three methods for visually representing information available within a personalized Web search system: tag clouds, term histograms, and term lists. These approaches not only describe the personalization model to the searcher, but also support interactive re-ranking of the search results. While preliminary evaluations did not find improvements beyond what was achieved with the personalization system, measures of subjective reactions showed an increased satisfaction in the search results. These results indicate that visual and interactive methods can be valuable for providing searchers with a sense of awareness and control during the search process.