异方差半变系数带有测量误差模型的经验似然

Empirical likelihood for semi-varying coefficient errors-in-variables heterocedastic models

考虑了伴有由外生变量函数和独立同分布的随机变量乘积组成的测量误差的异方差半参数变系数部分线性模型.分别在外生变量函数已知和未知的情形下,构造了模型相关参数向量的经验似然统计量,并证明其服从渐近卡方分布,即Wilks现象成立.

We were concerned with the semi-varying coefficient hererocedastic errors in variables,with the errors consisting of the exogenous variable functions and the i.i.d.variables.Specifically,we constructed an empirical likelihood ratio statistic for the parametric coefficient and proved the statistic to be asymptotically standard chi-squared distributed in view of the cases that the exogenous variable function was known or unknown.Thus the nonparametric Wilks phenomenon was derived.