Maximum likelihood estimation of functional relationships (Q1189470)

According to the author this monograph has two purposes: to explore the potentials of likelihood theory in the field of functional relationships and to summarize some classical results in this field. The second purpose is certainly met; relevant and most recent literature on the subject is briefly summarized and well classified, and integrated within the sections. As regards the first purpose, I still do not know the potentials except for some more information on the problems of consistency of likelihood estimators. The last property is here identified with correctness of the results. Mathematical rigour is not emphasized in favor of statistical insight. Nevertheless, e.g. liberate use of densities w.r.t. a differential element or to a measure, inverses or generalized inverses of matrices (even of ``generalized'' Moore-Penrose inverses) and similar items could have been avoidable. Also, various use of notions like observations, variables, variates, data, unobservables, latent values combined with dependent, explanatory, controlled, free or true makes understanding of the statistical phenomena not easy. At least three references in the text are not in the list at the end. Instead of continuing in this way my short summary: This is a good first draft of a manuscript on a relevant topic. I would like to read a careful and thorough revision.