Talk:Berndt–Hall–Hall–Hausman algorithm

Latest comment: 2 years ago by Superpronker in topic Usage section


Usage section

edit

The intro text correctly states that the procedure is only guaranteed to work for maximum likelihood models. Therefore, it seems bizarre that the Usage section starts by describing a general non-linear estimator. Moreover, the requirement that the criterion be on the sum form (an M-estimator) is a need-to-have, so that should be stated early on, not in the middle of the section. Superpronker (talk) 14:34, 3 November 2022 (UTC)Reply