Simulation and Application of Stock Composition Estimators

Publisher Website
Google Scholar
Abstract
Stock composition estimates for mixed populations may be obtained with a classification scheme and the results of classifying individuals from this mixed population and from the component subpopulations. Several estimators have been proposed. Simulation studies were conducted over a wide range of practical possibilities. These studies showed that the simplest estimator, together with a constraint, was more than adequate for practical application. Implementation of the constraint provides for a dramatic reduction in effort for scientists involved in stock identification work. Statistically valid, conservative confidence intervals were obtained with existing variance formulae.

This publication has 1 reference indexed in Scilit: