部分线性模型的一种新的异方差检验方法

A New Method of Heteroscedasticity Test in Partial Linear Models

主要考虑了部分线性模型的异方差检验问题。对部分线性模型的参数和非参数部分进行了估计,建立了对模型的随机误差进行异方差检验的经验似然比统计量,并证明该统计量渐近服从卡方分布。最后,通过数值模拟研究了统计量的有限样本性质。

The heteroscedasticity test of partial linear models is mainly considered. First, the parameter and non-parametric of the partial linear model are estimated. Then,the empirical likelihood ratio statistic for the heteroscedasticity test of this model is established,and it is proved that the statistic asymptotically has a limiting chi-square distribution. Finally,the finite sample properties of the statistic are studied by numerical simulation.