f-distributed errors

  1. S

    Linear regression - when errors are correlated and from different f distributions

    Hi all, This question relates to research assessing the hearing of new born babies. As mentioned in the title, I wished to solve the univariate model y = m*x + c + err, where each error term err_i comes from an F distribution, F(d1, d2), with d1 a known constant and d2 a known linear...