Evaluating the efficacy of clustered visualization in exploratory search tasks

Sarika S Kothari, Purdue University

Abstract

Information visualization has the potential to improve the quality of Web search results representation providing more context and novel ways to see relationships among items in a result set. The key objective of this research was to evaluate the potential of graphical visualization for representation of Web search results especially for exploratory search tasks. This is achieved by comparing the commonly used technique of ranked list representation of search results with the novel technique of representing these results using a cluster-based visualization technique. An experiment was designed in which participants performed Web searches for a set of predefined exploratory search scenarios. The number of links visited to complete each search task and the amount of time taken to complete the task was recorded. Participant feedback was collected to compare these two techniques. This information was then analyzed to evaluate efficiency of completing the search task, effectiveness at reaching the search goal, and user satisfaction with the two techniques. Important observations were made based on participant feedback on cluster-based visualization technique. This research study demonstrates the potential of cluster-based visualization techniques for Web search results representation as a complementary tool to currently available techniques to improve user experience for exploratory search tasks.

Degree

M.S.

Advisors

Mohler, Purdue University.

Subject Area

Educational technology|Computer science

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS