当模型误差不是白噪声时, 通常的估计方法无效,特别是当回归因子包含 后延变量时,估计不相合. 因此,文章研究了半参数可加测量误差模型的 白噪声检验. 提出了一个白噪声检验统计量, 并在模型误差是白噪声的零假设下证明了所提检验统计量服从渐近正态 分布. 随机模拟表明所提出的检验统计量具有良好的检验功效和水平.

When the model errors are not white noise, the popular estimation techniques become inefficient or invalid or inconsistent if the regressors contain lagged dependent variables. Therefore, in this paper we investigate the white noise semi-parametric additive measurement error models. We propose a white noise test statistic, and show that it follows an asymptotic normal distribution under the null hypothesis that the model errors are white noise. Simulation studies exhibit that the proposed test statistic has very good power and size