抽样估计中估计量的选择研究

Selection of Estimator in Sampling Estimation

经过研究证明V^YPR-V^YSR≤0,V^YPR≤V^YSR,这表明PPS抽样比率估计量比等概率抽样比率估计量有效。结合多方面分析可以得出结论:回归直线通过原点时,两种抽样方法的比率估计量与回归估计量抽样效率相等;当r>0.5时,PPS的抽样效率一般比等概率抽样有效;当b>a/X、并且r≥Cx/Cy时,等概率抽样的比率估计量比均值估计量有效。因而进行抽样调查时,一般是先估算r和a的值,当a值不等于零时,再估算b的值,最后根据上述的规则来决定选择采用哪种抽取样本的方法和抽样估计量。

It is proved by research that, which shows that the sampling rate estimator is more efficient than the equiprobability sampling rate estimator. Connected with an extensive analysis, it can be concluded that when the regression line gets across the origin, the rate estimator of two sampling methods is equal to the sampling efficiency of regression estimator, when , the sampling efficiency otis more efficient than the equiprobability sampling, and when and meanwhile , the rate estimator of equiprobability sampling is more efficient than the mean estimator. Therefore, when conducting the sampling research, we generally estimate the numerical value of and first. If the numerical value of is not equal to zero, we calculate the numerical value of then. Finally, based on the rules mentioned above, we decide on which sampling method and sampling estimate should be selected.