Supporting Data Mining of Large Databases by Visual Feedback Queries

In this paper, we describe a query system that provides visualrelevance feedback in querying large databases. Ourgoal is to support the process of data mining by representingas many data items as possible on the display. By arrangingand coloring the data items as pixels according to their relevancefor the query, the user gets a visual impression of theresulting data set. Using an interactive query interface, theuser may change the query dynamically and receives immediatefeedback by the visual representation of the resultingdata set. Furthermore, by using multiple windows for differentparts of a complex query, the user gets visual feedbackfor each part of the query and, therefore, may easier understandthe overall result. Our system allows to represent thelargest amount of data that can be visualized on current displaytechnology, provides valuable feedback in querying thedatabase, and allows the user to find results which, otherwise,would remain hidden in the database.

Authors: Keim D. A., Kriegel H.-P., Seidl T.
Published in: Proc. IEEE 10th Int. Conf. on Data Engineering (ICDE 1994), Houston, Texas
Publisher: IEEE Computer Society - Washington,USA
Language: EN
Year: 1994
Pages: 302-313
ISBN: 0-8186-5400-7
Conference: ICDE
Files:
Type: Conference papers (peer reviewed)
Research topic: Exploration of Multimedia Databases