右删失时间序列AR模型参数估计  

Parameters Estimation Based on AR Time Series Model with Right-censored Data

作　　者：杨琦[1] 王晓慧[1] 曹显兵[1] YANG Qi WANG Xiao-hui CAO Xian-bing(School of Science, Beijing Technology and Business University, Beijing 100048, China)

机构地区：[1]北京工商大学理学院,北京100048

出　　处：《数学的实践与认识》2017年第10期209-214,共6页Mathematics in Practice and Theory

基　　金：2016年研究生科研能力提升计划项目资助

摘　　要：通过Kaplan-Meier估计和Nelson-Aalen估计得到了平稳时间序列被另一平稳序列右删失下.AR模型的参数估计.首先,通过与完全数据下的参数估计进行对比,说明了两种估计方法的效果.然后,根据计算机模拟的样本量以及删失率的不同,对比了两种估计的优劣,并且模拟结果表明两种估计是有效的.Abstract： K-M estimation and N-A estimation are used to estimate parameters of AR model which is right-censored by another time series in this paper. Firstly, the effectiveness of two methods are explained by the comparison with complete data parameter estimation. Secondly, with different sample size and censored percentage in simulation, we contrast the advantages and disadvantages between these two methods, and the simulation results show that two estimation methods are effective.

关 键 词：右删失数据 K—M估计 N—A估计 AR模型 R软件 

分 类 号：O212.1[理学—概率论与数理统计]