权数的评估和控制研究
发布时间：2019-08-01 07:35
【摘要】：用样本推断总体,权数的作用十分重要。通过权数,不仅能将样本还原到总体,还能调整样本的结构,从而提高估计的精度。但权数也是把双刃剑,如果调整后的权数差异过大有可能增大样本量的方差。本文从权数控制的角度出发,利用权效应对主观截取和计算截取两种权数控制方法进行对比研究。最后得出,主观截取对权数的控制要稍优于计算截取的效率,并且得出若采用主观截取、计算截取的方法,?(28)0.05、?(28)0.95可分别作为一般的分割点。
[Abstract]:It is very important to infer the population from the sample. Through the weight, the sample can not only be reduced to the population, but also the structure of the sample can be adjusted, so as to improve the accuracy of the estimation. However, the weight is also a double-edged sword, if the adjusted weight difference is too large, it is possible to increase the variance of the sample size. From the point of view of weight control, this paper makes a comparative study on subjective interception and computational interception by using weight effect. Finally, it is concluded that the control of the weight of subjective interception is slightly better than that of calculating interception, and it is concluded that if subjective interception is used, the method of subjective interception can be used as the general dividing point of? (28) 0.05 and? (28) 0.95 respectively.
【作者单位】： 中国人民大学统计学院;中国人民大学应用统计科学研究中心;湖北经济学院;
【分类号】：F224
，
本文编号：2521622
[Abstract]:It is very important to infer the population from the sample. Through the weight, the sample can not only be reduced to the population, but also the structure of the sample can be adjusted, so as to improve the accuracy of the estimation. However, the weight is also a double-edged sword, if the adjusted weight difference is too large, it is possible to increase the variance of the sample size. From the point of view of weight control, this paper makes a comparative study on subjective interception and computational interception by using weight effect. Finally, it is concluded that the control of the weight of subjective interception is slightly better than that of calculating interception, and it is concluded that if subjective interception is used, the method of subjective interception can be used as the general dividing point of? (28) 0.05 and? (28) 0.95 respectively.
【作者单位】： 中国人民大学统计学院;中国人民大学应用统计科学研究中心;湖北经济学院;
【分类号】：F224
，
本文编号：2521622
本文链接：https://www.wllwen.com/jingjifazhanlunwen/2521622.html
