Bias of LS estimators in nonlinear regression models with constraints. Part II: Biadditive Models

摘要: General results giving approximate bias for nonlinear models with constrained parameters are applied to bilinear in ANOVA framework, called biadditive models. Known on the information matrix and asymptotic variance of summarized, Jacobians Hessians response constraints derived. These intermediate basis any subsequent second order study model. Despite large number involved, formulae turn out be quite simple due orthogonal structure In particular, estimators shown approximately unbiased. Some simulations assess validity approximations.