ContactPerson: agarg@cse.buffalo.edu ### Begin Citation ### Do not delete this line ### %R 2001-03 %U /web/faculty/agarg/tmp/ivfdc2-tr.ps %A Bhadra, Debangshu %A Garg, Ashim %T An Interactive Visual Framework for Detecting Clusters of a Multidimensional Dataset %D March 27, 2001 %I Department of Computer Science and Engineering, SUNY Buffalo %K visualization, data mining, cluster, databases, interactive, animation, graphics %X A key problem in data mining applications is that of detecting clusters, i.e., groups of closely related data items, of a multidimensional dataset. Users are adept at detecting clusters in visually presented information. We, therefore, present an interactive framework that allows a user to visualize a multidimensional dataset, and detect clusters in it by repeatedly changing its visualization using interactive operations such as parameter re-adjustment, zooming, subset selection, and cluster probing. Our experiments also suggest that animation provides important clues in detecting clusters. -------------------------------------------- Managed by Request Tracker