部分线性单指标模型的序列相关性检验

Testing Serial Correlation in Partial Linear Single-index Model

部分线性单指标模型是一类重要的半参数统计模型,对于模型中的误差,如果存在序列相关,那么在进行统计推断时会遇到诸多问题。针对该情况,对部分线性单指标模型的序列相关性进行检验,用VT,P检验的方法构造VT,P检验统计量,得到了零假设下检验统计量的渐进分布。数值模拟验证结果表明,该检验有较好的检验功效。

Partial liner single-index model is an important semiparametric statistic model. We know that if serial correlation exists for errors of regression models, then many problems will appear in the procedure of statistical inference. This paper is mainly to test serial correlation in partial liner singleindex model. We use the method of Vr.e test to construct testing statistics, and further derive the as- ymptotic distribution under null hypothesis, Simulation results show that our test has good effect.