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.
出处
《兰州理工大学学报》
CAS
北大核心
2014年第6期160-164,共5页
Journal of Lanzhou University of Technology
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
empirical likelihood
confidence region
heteroscedasticity
varying-coefficient partially linearmodels
measurement error