Composite Estimation of Totals for Livestock Surveys
美国农业部国家农业统计服务局使用四种初步估计量来估算主要州的牲畜存栏量，本文提出一种复合估计量（加权平均）来整合它们，并比较了固定权重和自适应权重两种类型，通过杰克刀法评估方差和均方误差，1984年数据支持其应用。
Abstract Abstract Four different preliminary estimators are employed by the National Agricultural Statistics Service of the U.S. Department of Agriculture to obtain the final estimate of livestock inventories of major states. To combine these four estimators a composite estimator (which is a weighted average) is proposed. Two types of composite estimators—one with fixed weights, the other with adaptive weights—are investigated. The composite estimator with adaptive weights is derived by minimizing a quadratic function subject to linear constraints. The variance and mean squared error of the composite estimator with adaptive weights are evaluated by the jackknife method. Numerical results based on the data from the 1984 June Enumerative Survey conducted by the Department of Agriculture support the use of composite estimation. Key Words: Common meanConvex programmingDomain estimationJackknifeMultiple-frame methodologyStratified sampling