The Estimation of Many Parameters

Publisher Website
Google Scholar
Abstract
This article contains a proof of the theorem "In an infinite universe there are k frequency grades, k being arbitrary, the probability for the ith grade being pi. The pi are functions of known form of h parameters 01, O2; . . ., 0h of which the first derivatives all exist. In indefinitely large random samples the estimates of the parameters derived from the maximum likelihood equations are those which minimize the generalized variance." The proof is based upon Wilks'' generalized variance and the theory of compound matrices.

This publication has 0 references indexed in Scilit: