基于平衡轮换样本调查的时间序列建模
发布时间：2019-01-27 22:15
【摘要】：连续性抽样调查由于能够描述目标总体随时间的动态变化过程,吸引了越来越多国内外学者的关注。国外连续性抽样的研究已经十分成熟,在已知的轮换模式下,建立合适的模型,使得模型能较好地描述数据的真实生成过程,从而得到精度更高的目标估计量。文章建立一般轮换模式r~m_1~r~(m- 1)_2下的时间序列模型,然后以6~3~6~2模式为例,利用状态空间模型和卡尔曼滤波,给出已有信息下的最优估计,有效减少抽样误差,提高样本的估计精度。
[Abstract]:Continuous sampling survey has attracted more and more scholars' attention because of its ability to describe the dynamic process of target population over time. The research of continuous sampling abroad has been very mature. Under the known rotation mode, a suitable model is established, which can better describe the real process of data generation and obtain a more accurate target estimation. In this paper, the time series model of the general rotation model rnm-1 _ 2 is established. Then, taking the 6 ~ 3 ~ (-1) 2 model as an example, using the state space model and Kalman filter, the optimal estimation under the available information is given, which can effectively reduce the sampling error. Improve the estimation accuracy of the sample.
【作者单位】： 暨南大学经济学院;
【基金】：全国统计科学研究计划项目(2012LY014) 霍英东教育基金会项目(141096) 广东省优秀博士学位论文资助项目(sybzzxm201120)
【分类号】：F224
[Abstract]:Continuous sampling survey has attracted more and more scholars' attention because of its ability to describe the dynamic process of target population over time. The research of continuous sampling abroad has been very mature. Under the known rotation mode, a suitable model is established, which can better describe the real process of data generation and obtain a more accurate target estimation. In this paper, the time series model of the general rotation model rnm-1 _ 2 is established. Then, taking the 6 ~ 3 ~ (-1) 2 model as an example, using the state space model and Kalman filter, the optimal estimation under the available information is given, which can effectively reduce the sampling error. Improve the estimation accuracy of the sample.
【作者单位】： 暨南大学经济学院;
【基金】：全国统计科学研究计划项目(2012LY014) 霍英东教育基金会项目(141096) 广东省优秀博士学位论文资助项目(sybzzxm201120)
【分类号】：F224
【相似文献】
相关期刊论文 前8条
1 李小胜;时间序列建模过程应注意的几个问题[J];统计与决策;2003年06期
2 张学斌,刘嘉q ,刘菁,刘泊e，
本文编号：2416752
本文链接：https://www.wllwen.com/jingjifazhanlunwen/2416752.html
