Grant: Discovery Grants Program - Individual

In the past few decades, the field of robust statistics as a well-defined research area has experienced substantial growth in general and expanding in promising new directions in particular. The robust approach to designs of statistical experiments and statistical inferences aims at deriving reliable statistical procedures which are not sensitive to small departures from model assumptions or small proportions of contamination in data. This research program proposal involves four major research projects in the following three coherent directions: 1) theoretical development of optimal and robust designs for generalized linear fixed models and generalized linear mixed models; 2) robust designs, with applications in accelerated life testing, incorporating both censoring and stress-loading schemes; and 3) robust estimation for autoregressive models with non-constant conditional variances and autoregressive moving average models with non-constant conditional variances.