异方差半参数变系数部分线性EV模型的经验似然  

Empirical likelihood of semiparametric,varying-coefficient,heteroscedastic,and partially linear EV models

作　　者：夏亚峰[1] 达虎 

机构地区：[1]兰州理工大学理学院,甘肃兰州730050

出　　处：《兰州理工大学学报》2014年第6期160-164,共5页Journal of Lanzhou University of Technology

摘　　要：利用经验似然方法考虑异方差半参数变系数部分线性EV模型中兴趣参数置信域的构造.分别在误差方差已知和未知情形下,构造模型参数部分的经验对数似然比统计量,并验证非参数情形下的Wilks定理成立.模拟研究表明经验似然估计方法具有优良的性质.The empirical likelihood method is used to study the confidence regions construction ot the in- terested parameters in a semiparametric, varying-coefficient, heteroscedastic, and partially linear EV mod- els. The statistic quantities of empirical log-likelihood ratio in parametric part of the model are constructed respectively for the case of known and unknown error variance, and the nonparametrie version of Wilks＇ theorem is verified to hold Simulation study shows that the empirical likelihood estimate approach has nice properties.

关 键 词：经验似然 置信域 异方差 变系数部分线性模型 测量误差 

分 类 号：O212.7[理学—概率论与数理统计]