Heteroscedastic additive models - Estimating the fixed effects and covariance matrix parameters

This work aims to deduce estimators for the unknown parameters of fixed effects and covariance matrix structure in heteroscedastic additive design. In order do that, design will be projected onto orthogonal complement subspace spanned by columns effects, Kronecker product used produced unbiased matrix, then such produce an estimator effect vector. Moreover, coefficient determination both derived. A simulation study conducted, a numerical example explored.