Scott 9708176 This proposal requests funding to support young researchers and speakers to attend the 29th symposium on the Interface of Computing Science and Statistics in Houston, May 14-17, 1997, along with related conference expenses. The theme of this year's interface is "Mining and Modeling Massive Data Sets In Science, Engineering, and Business," which is related to other special NSF initiatives. The Symposium is designed to bring together top researchers and workers in several related disciplines to focus on a breaking theme, and to bring younger researchers, graduate students, and interested scientifically minded individuals from a wide cross-section of fields and industries to interact and exchange information and ideas. This year's program has drawn a field of invited speakers from an international group, as well as some of the best minds in the American academic, computer, and business worlds. Furthermore, a large and diverse group of young individuals will attend. The conference also has a sub-theme of the environment and quantitative environmental science. The conference is sponsored by the Interface Foundation of North America, a non-profit educational corporation. The Statistics Department at Rice University is hosting the meeting with David W. Scott as organizer and program chair. Co-hosting institutions include M.D. Anderson Cancer Center and U. Texas School of Public Health. The Symposium has a number of cooperating societies. Data mining holds the promise of dramatically improving our utilization of massive data records accumulating in all areas of government, industry, business, and research. Some of the best statisticians, computer scientists, federal researchers, businessmen, and others are contributing to providing new algorithms and ideas for finding structure in massive data sets. Over 27 invited and 12 contributed paper sessions will be offered. Keynote speaker Jerry Friedman of Stanford will discuss the role of statistics in data mining research. This work poses serious hardware and software problems for computer vendors, and speakers from IBM, SGI, Oracle, AMS, and others will discuss hardware and software issues related to massive data sets management and analysis. Other sessions will describe case studies and war stories. A number of marketing researchers have won permission to describe usually proprietary work that should have educational scientific value. Special luncheon presentations will offer a more informal atmosphere and interaction with leading researchers in computing, genetic and molecular modeling, and network understanding. Visualization is a common thread, with sessions on the uses of virtual reality, exploration, and spatial statistics for massive data sets. The CEO's and chief scientists of some of the leading data mining and statistical software houses will be speaking at this year's meeting. Large scale optimization and wavelet techniques will also be covered among the many sessions. Each speaker will provide a manuscript for a proceedings of the conference, which is to be published in the early Fall. Thus the results and information provided during the session will be available to a very wide audience at low cost.
