scientific article; zbMATH DE number 3936267

From MaRDI portal
Publication:3707161





[For the entire collection see Zbl 0564.00017.] Consider the model E(\b{Y})\(=X{\underline \beta}\) where \b{Y} is an \(n\times 1\) vector of observations, \(y_ i\), taken at the design points, \b{x}\({}_ i=(x_{i1},x_{i2},...,x_{ik})\), \(i=1,2,...,n\), in a k- dimensional space, X is an \(n\times p\) matrix of explanatory variables, \({\underline \beta}\) is a \(p\times 1\) vector of unknown parameters and var(\b{Y})\(=V\sigma^ 2\), V known and \(\sigma^ 2\) being the experimental error variance. In the investigation of the model, there is an extensive literature on the relationship between the ordinary least squares estimator and weighted least squares estimator, i.e., the situation in which the observations have an arbitrary covariance matrix. In the present paper, it is shown that when the weighted and unweighted least squares estimators are not equal, a weighted Lagrangian interpolation formula is obtained.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3707161)