Prediction in Generalized Least Squares

Wm. Craig Riddell · The American Statistician · 1977

An alternative derivation of the best linear unbiased predictor for the general linear model (in which the errors are nonspherical) is provided. This derivation is simpler than the original version due to Goldberger [3], and may be preferable for instructional purposes.

Read the paper · More papers on PaperTik