This paper deals with the problem of P-optimal robust designs for multiresponse approximately linear regression models. Each response is assumed to be only approximately linear in the regressors, and the bias function...This paper deals with the problem of P-optimal robust designs for multiresponse approximately linear regression models. Each response is assumed to be only approximately linear in the regressors, and the bias function varies over a given L2--neighbourhood. A kind of bivariate models with two responses is taken as an example to illustrate how to get the expression of the design measure.展开更多
基金National Natural Science Foundation of China(10671129)the Special Funds for Doctoral Authorities of Education Ministry of China(20060270002).
文摘This paper deals with the problem of P-optimal robust designs for multiresponse approximately linear regression models. Each response is assumed to be only approximately linear in the regressors, and the bias function varies over a given L2--neighbourhood. A kind of bivariate models with two responses is taken as an example to illustrate how to get the expression of the design measure.
