The investigator will develop statistical methods for testing for patterns in multilinear models, methods applicable even when the parameters of the patterned model cannot be identified. Statistical methodology for incorporating nonparametric smoothing and covariates when several vectors are measured on each subject will also be developed. Detailed real examples will be analyzed to illustrate the methods and demonstrate their applicability. The fitting of a low-rank model to a three-way data array containing noise is an active area of research with important applications in chemometrics and psychometrics. The project furthers VPW program objectives to provide opportunities for women to advance their careers in science or engineering through research, and to encourage other women to pursue careers in these areas through the investigators' enhanced visibility as role models on the host campuses. In this project, the proposed activities which contribute to the second objective include: actively participating in and giving lectures for the Biostatistics Workshop (a joint enterprise between the Department of Statistics and the School of Medicine which meets twice weekly during the Fall Quarter); giving lectures on her research to the Statistics Department's faculty and graduate students as part of a biweekly lecture series; participating in the Department's program of collaborative research with active research scientists from around the University; and meeting with undergraduate majors to discuss career opportunities in Statistics.
