基于3种估计量的冬小麦播种面积 总体外推效率比较

COMPARISONS ON EXTRAPOLATION EFFICIENCY OF MULTIPLE ESTIMATORS FOR WINTER WHEAT ACREAGE ESTIMATION

关键词:  估计量 抽样调查 播种面积 外推总体 误差估计

Key words:  estimator  survey sampling  sown area  extrapolation  error analysis

摘要:

为进一步改善现行农作物播种面积抽样外推总体的精度和稳定性，该文以安徽省蒙城县为研究 区，选取冬小麦播种面积为研究对象，采用简单随机抽样方法进行样本抽选，基于样本的地面调查与遥感 调查冬小麦播种面积数据构建3种统计估计量（简单估计量、比率估计量和回归估计量），选取相对误差 和变异系数CV为外推总体效率评价指标，对3种估计量外推研究区冬小麦播种面积总体的效率进行了定 量比较，结果表明:样本内的冬小麦播种面积地面调查精度高于遥感调查，但两种调查方式下的样本观测 值间差异不显著;样本的地面调查与遥感调查结果间存在极显著的正比例关系;3种估计量中，比率估计 量外推总体的效率最高（相对误差和CV值最小），其次是回归估计量，简单估计量外推总体的效率最低。 该文可为优选农作物面积空间抽样调查统计估计量提供参考依据。

Abstract:

The experiments on quantitatively comparing the extrapolation efficiency of multiple estimators were con- ducted to improve the accuracy and stability of current survey sampling for crop sown area estimation. The winter wheat sown area in Mengcheng County of Anhui provinces selected as study object， and the simple random sampling method is used to draw the samples for estimating population values in this experiment. Samples observations ob- tained by ground survey and remote sensing survey are employed to construct three kinds of estimators（simple esti- mator， ratio estimator and regression estimator）. In order to evaluate the inference efficiency of different estimators， the relative errors and coefficients of variation are selected as the criterion of extrapolation efficiency. The experi- mental results demonstrated that: the precision of samples ground survey was higher than that of remote sensing in- vestigation， but there were no obvious differences between the two survey results of winter wheat acreage; There was a significant direct proportional relationship between the two sample survey results; The extrapolation efficiency of ratio estimator was the best， followed by regression estimator， the simple estimator was the worst among the three kinds of estimators. This paper can provide a solution for optimizing estimators to extrapolate crop acreage at region- al level.