结构型EV模型中的参数估计-概率论与数理统计专业论文.docxVIP

中文摘要本文利f}j霉星壹鱼谨的方法，分删给出了简荦毪垡笙毽型i：：!蟹和一股线性 中文摘要 本文利f}j霉星壹鱼谨的方法，分删给出了简荦毪垡笙毽型i：：!蟹和一股线性 结构型EV模型中的参数估计，并讨论了估计的强相合性与渐近『F念性。得到了 ___’‘‘。■^_”_’—’———_一 911】中对函数型EV模型鲍磷究襁平行的结暴。 对予简单线性结构型EV模型，即 舌f嚣x，+正，，叮。／=Y。+su=口十_，∥+sp，J=l，’·‘n，；f=l，。一，七， (0．1) ‘蒸中岛，％。毛，毛分鬟是蠢窝爨t程霾变爨只靛蕊溺{耋鞠鼹测误差户／ 、 受【l】中对函数型EV的研究的启发，本文用矩方法给出了参数卢，口的估计。 在撼样以及自变量秘瑷测误差满懋遗当条传下，我们剽用“烟权独立随机变鬟程 收敛性”的有关定理，得到了估计的强相合性。当测量误差臌工人正态分布时，运 用特征函数这一工具得到了估计的渐近正态性；当测量误差不服从正态分布时， 只要测爨谖差独立，曩叁变量懿残溅误差(疋)分毒对称，狂在2+r(荬r为任 意正数)阶有限矩，则利用Lyapunov它：坠拯腿畿理即可得到估计的渐近正态性。 对于一般线性结构型Ev模型，阳p 弋 岛=菇，+《气，弩#=Y。+岛=搿+x；p÷￡口，歹=l，’-‘H，；}=l， ，意， (o．2) (其中乞，％，吼，气分别是自变量爿，和因变量M的观测值和观测误差， 岛，五，磊是p缝淀爨壮文给撼了参数多，聪浆售诗，著凌逶当条传下，褥裂 了估计的强相合性。同时，利用【1】中对函数型EV模型的研究成果，运用取条件 期望的方法，也得到了估计的渐避难态性。 AbstractBy Abstract By taking repetitive observations in this paper,parametric estimators are obtained respectively in a simple structural EV linear model and a linear structural EV model with vector explanatory variables．The strong consistency and asymptotic normality of the estimators are also discussed．All the results are parallel with those in functional EV models of[1]． The simple linear structural EV model studied in the paper can be written as 亏p=工，+氏，玎．，=Y。+g¨=d十x，∥+占口，／=1，’·’n，；f=l，+1。，k，’‘ (o．1) Where{U，r／U，6U，EU are the observation and measurement error of x?and Yi respectively．With the inspiration of the research of functional EV models in 【l】，the estimators of the parameters of爹，aare obtained by using the method of moments．When the sampling，argumem and measurement eITor are subject to some conditions including毛and嘞are i，i．d respectively，some theorems about weighted sums of independent random variables are used to prove the strong consistency of the estimators．When the measurement errors are subject to normal distribution，characteristic functions are used to prove the asymptotic normality of the estimators．Otherwise，the property carl be proved by means of Lyapunov central limit theorem，as long as 6{and￡q are independent，the distribution of 6{is symmetric，and there are