The numerical evaluation of the maximum-likelihood estimate of mixture proportions

Abstract

Let p1,…,pm be multivariate density functions and let α = (α1,…,αm)T be a proportion vector defining the mixture density , where and αi ≥ 0 for i = l,…,m. This paper investigates the problem of evaluation of the maximum-likelihood estimate of α. An acreage estimation application is presented using remotely sensed data.