Abstract
This dissertation is an attempt to provide analytic properties of a special nonlinear estimation which arose in the analysis of earth resource technological satellite (ERTS) data. The problem is concerned with the estimation of crop proportions from observed multivariate frequencies of so-called channel intensities. A variate transformation model is employed to estimate the crop proportions. The procedure is based on a transformation of a mixed distribution of crop "signature" into their known marginal distributions. The parameter estimates for the transformation are obtained simultaneously with the proportion of each crop type of interest on the land segment. Work has been done to show that at every point of cumulation of the sequence of parameter vectors generated by the estimation procedure the Kuhn-Tucker conditions are satisfied. A sufficient condition for the existence of a unique point of cumulation of the sequence of parameter vectors has been provided.
Al-Khalidi, Abdul Sattar (1978). An analytic study of a minimum x²-estimation procedure. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -776825.