Abstract

A critical aspect of developing an informatics based design of new materials and processes is to explore existing data sets (experimental or simulated) and to search for physically meaningful associations. Classical statistical methodologies may prove to have limited value when data sets are not large and/or when there can be significant uncertainty in the data sets. In this presentation we present examples of data analysis approaches in materials science based on fuzzy clustering and other techniques. The value of such techniques are discussed in the context of developing reliability analysis of materials design problems.

Biography

Krishna Rajan holds joint appoints in the School of Engineering as Professor of Materials Science and in the Faculty of Information Technology program as Professor of Materials Informatics at Rensselaer Polytechnic Institute. Professor Rajan received the Sc.D. degree in 1978 from the Massachusetts Institute of Technology in Materials Science. He was a Research Fellow at the Metallurgy and Materials Science Department at Cambridge University until 1980. From 1980-87 he was staff scientist at the National Research Council of Canada. In 1987, he joined the faculty at Rensselaer being appointed as a full professor in 1993 . Professor Rajan has published over 160 publications, edited seven monographs and presented over 170 scientific lectures nationally and internationally. He has received the Lilly Endowment Teaching Fellowship, AMAX Foundation Research Fellowship, the ASM International Visiting Lecture Award and the NRC /NAE E.Europe Fellowship. He has held numerous visiting appointments internationally including the Max-Planck Institute, Stuttgart, Germany; University of Auckland, New Zealand; University of Wollongong, Australia and the Slovak Academy of Sciences. Professor Rajan also serves as the editor for Materials Science and Engineering for the McGraw-Hill Encylopedia of Science and Technology and Yearbook. He is also was a panel member of the recently commissioned Department of Energy report on the "Future Information Infrastructure for the Physical Sciences" Dr. Rajan research interests focus on the microstructural aspects of materials science with a major effort in the applications of high resolution electron microscopy, however he has extended his research into coupling new developments in computer science and mathematical research into combinatorial materials science and informatics.. Based on these efforts he has established the first academic program in Materials Informatics in a major Materials Science department in the United States.

Michael J. Doyle is Director of the Formulations Consortium at Molecular Simulations Inc. San Diego CA. He received his B.A. , M.A. and Ph.D in Natural Sciences (Statistics, Chemistry and Biochemistry) from Cambridge University, UK , where he investigated new applications of crystallographic data to chemistry and molecular modeling. Areas covered include topological analysis of rings systems and systematic packing analysis of quinoic acids. He also worked on software development for the Cambridge Crystallographic Database, and was responsible for all the statistical, cluster and conformational analysis tools within that software environment, and directed the departmental single crystal X-ray characterization facility for two years. His industrial experience includes postdoctoral work at ICI Agrochemicals (Jellot's Hill, UK). Before moving to MSI, Michael worked at BP Research as a research scientist, responsible for scale, crystal growth, and corrosion projects as well as participation in zeolite and other materials research areas.

1.	Krishna Rajan Professor of Materials Engineering Materials Research Center Science and Engineering Rensselaer Polytechnic Institute Troy, NY 12180 Tel. (518) 276-6126 email: Rajank@Rpi.Edu

2.	Michael Doyle Director of the Formulations Consortium Molecular Simulations Inc. 9685 Scranton Road San Diego, CA 92121-3752 (858) 799-5711 Tel. 619-799-5358