摘要
考虑随机系数自回归模型yt=Φtyt-1+ut,其中Φt为随机系数,ut为随机误差。在允许Φt与ut相依以及Εu^4t无穷的条件下,构造了误差方差的自加权估计,并证明了该估计的渐近正态性。最后通过数值模拟,说明自加权估计的稳健和有效性。
Consider the random coefficient autoregressive model yt =Φt yt - 1 + u t, in which the random coefficients are permitted to be correlated with the random errors. A robust self-weighted M-estimator of the error variance is proposed, and shown to be asymptotically normal with Eu^4t being possibly infinite. Some simulation studies are also given to show the good performance of the self-weighted estimator.
作者
傅可昂
丁丽
李婷
陈豪
何文凯
FU Keang;DING Li;LI Ting;CHEN Hao;HE Wenkai(School of Statistics and Mathematics,Zhejiang GongshangUniversity,Hangzhou 310018,China)
出处
《浙江大学学报(理学版)》
CAS
CSCD
北大核心
2019年第4期416-421,共6页
Journal of Zhejiang University(Science Edition)
基金
浙江省自然科学基金资助项目(LY17A01004)
教育部人文社科研究青年基金项目(17YJC910002)
浙江省一流学科A类项目(浙江工商大学统计学)
浙江工商大学研究生科研创新基金项目
关键词
广义随机系数自回归
误差方差
自加权估计
渐近正态
generalized random coefficient autoregression
error variance
self-weighted estimation
asymptotic normality