ASA Statistical Graphics
Session Slot: 10:30-12:20 Thursday
Estimated Audience Size: 200
AudioVisual Request: VCR, slide projector
Session Title: Graphics for Data Mining
Data mining as both a technology and a buzz word has become the darling
of the computer science and business applications world. At first sight,
this would seem to be nothing more than a repackaging of well-known
techniques of exploratory data analysis. However, requirements of data
set size and complexity often make traditional exploratory graphical
techniques unusable. New approaches which can simultaneously deal with
massive data sets, high dimensionality and missing data are requirements
of new graphical techniques. This session explores some of these
directions.
Theme Session: Yes
Applied Session: No
Session Organizer: Wegman, Edward J. George Mason University
Address: Center for Computational Statistics George Mason University, MS 4A7 Fairfax, VA 22030-4444
Phone: 703-993-1691
Fax: 703-993-1700
Email: ewegman@gmu.edu
Session Timing: 110 minutes total (Sorry about format):
110 minutes total First Speaker - 30 minutes Second Speaker - 30 minutes Third Speaker - 30 minutes Discussant - 10 minutes (or none) Floor Discussion - 10 minutes
Session Chair: Symanzik, Juergan George Mason University
Address: Center for Computational Statistics MS 4A7 George Mason University 4400 University Drive Fairfax, VA 22030-4444
Phone: (703) 993-3786
Fax: (703) 993-1700
Email: symanzik@galaxy.gmu.edu
1. A Successful System Using Data Mining for Direct Marketing Response Prediction
Wendelberger, James G., Consultant
Address: 620 Los Pueblos Los Alamos, NM 87544-2615
Phone: 505 662 2838
Fax: 505 662 2838
Email: jimwen@sisna.com
Abstract: The "GAIN" system to accomplish data mining for marketing response data is described. The system was utilized in the "Knowledge and Data Discovery Cup '97 Competition" to tie for first place. The system is utilized for direct marketing response prediction. Practical aspects of the creation of the system are described. Problems which have been solved are presented. Research areas are described.
2. Visualising Large Datasets: Issues and Ideas
Stone, Glenn, CSIRO Mathematical and Information Sciences, Australia
Address: CSIRO Mathematical and Information Sciences, Locked Bag 17, North Ryde, NSW 2113, Australia
Phone: +61 2 9325 3100
Fax: +61 2 9325 3200
Email: Glenn.Stone@cmis.csiro.au
Lovell, David, CSIRO Mathematical and Information Sciences, Australia
Abstract: In the typical application areas of data mining, for example finance and insurance, datasets can consist of millions of observations and hundreds of records. This has significant implications for traditional visualisation methods such as scatter plots and scatter plot matrices, coplots and 3D rotating point clouds. We consider techniques applicable to the data mining problems of exploratory data analysis, market segmentation and identifying unusual groups.
3. Crystal Vision: A Graphical Tool for Data Mining
Luo, Qiang, George Mason University
Address: Center for Computational Statistics George Mason University, MS 4A7 4400 University Drive, Fairfax, VA 22030-4444
Phone: (703) 993-3648
Fax: (703) 993-1700
Email: qluo@jupiter.galaxy.gmu.edu
Abstract: Crystal Vision is an extension of the SGI based software known as ExplorN, the latter developed as a research demonstration tool at George Mason University. The intent is to have a high performance software operating on a UNIX workstation with the capability of real-time or near real-time graphical analysis of large, high dimensional data sets. The software is capable of handling 20 to 30 dimensional data and it has been tested with 250,000 observations. This talk will discuss both the computational considerations taking advantage of some of the architectural features of the Silicon Graphics workstation as well as some of the statistical and graphical considerations.
Discussant: Poston, Wendy L. Office of Naval Research
Address: Office of Naval Research Division of Mathematical and Computer Sciences 800 N. Quincy Street, Arlington, VA 22217-5000
Phone: (703) 696-4320
Fax: (703) 696-2611
Email: postonw@onr.navy.mil
List of speakers who are nonmembers: None