FGLS

In statistics we have a trick, To estimate parameters quick. It's called GLS, a method grand, For linear regression it's in demand. When residuals have correlation, GLS comes to the equation. It improves efficiency, reduces risk, And helps us make inferences brisk. In 1935, Aitken was the first, To describe GLS, we must rehearse. But to use it, we need to know, The covariance matrix, high and low. If we don't have that knowledge yet, FGLS is the method we should get. Feasible Generalized Least Squares it's called, But guarantees of improvement may be stalled. So when you're doing regression math, Remember GLS on your chosen path. It helps us estimate with precision, And makes statistics a delightful mission. Random page: JOE DIMAGGIO