Search

The problem of estimating the parameters in a finite mixture is considered. The approach is based on an integral equation formulation of the form h sub t (x) = integral (limits b and a) f(x-y) g sut t (y)dy where h sub t is a smoothed version of h and g sub t is a prior function that tends to be concentrated on the translation values. A solution for g sub t that uses the method of regularization and one based on a posterior operator approach is considered. Numerical simulations are presented to bring out some of the estimation and numerical problems of these approaches.

Studies completed in fiscal year 1981 in support of the clustering/classification and preprocessing activities of the Domestic Crops and Land Cover project. The theme throughout the study was the improvement of subanalysis district (usually county level) crop hectarage estimates, as reflected in the following three objectives: (1) to evaluate the current U.S. Department of Agriculture Statistical Reporting Service regression approach to crop area estimation as applied to the problem of obtaining subanalysis district estimates; (2) to develop and test alternative approaches to subanalysis district estimation; and (3) to develop and test preprocessing techniques for use in improving subanalysis district estimates.

The results of the Domestic Crops and Land Cover Classification and Clustering study on large area crop estimation using LANDSAT and ground truth data are reported. The current crop area estimation approach of the Economics and Statistics Service of the U.S. Department of Agriculture was evaluated in terms of the factors that are likely to influence the bias and variance of the estimator. Also, alternative procedures involving replacements for the clustering algorithm, the classifier, or the regression model used in the original U.S. Department of Agriculture procedures were investigated.

The performance of the USDAs EDITOR system is evaluated. The system processes Landsat imagery and estimates crop hectarage for large areas based on a regression estimator developed on a sample with known ground truth. It is found that use of multitemporal data over unitemporal significantly improves the hectarage estimates, and a 15% reduction in the r-squared of the regression occurs when independent and jackknifed test sets are used to evaluate the performance of the estimator. When an alternative clustering algorithm, CLASSY, is substituted for the current EDITOR method, estimator performance is improved with reduced need for analyst decisions. It is recommended that the CLASSY clustering algorithm and some form of jackknifing be implemented on EDITOR.

There are no author-identified significant results in this report.

There are not author-identified significant results in this report.

There are no author-identified significant results in this report.

There are no author-identified significant results in this report.