This paper introduces a robust frequency domain empirical likelihood inference procedure for the parametric component in the spectral densities of stationary processes. We construct the empirical likelihood function by using a new spectral estimating function to achieve robustness against contamination in the spectral density. Simulation studies demonstrate the good performance of the proposed robust frequency domain empirical likelihood method, which produces more accurate confidence regions than the ordinary empirical likelihood counterpart.

中文翻译：

---

时间序列的稳健经验似然

本文介绍了一种稳健的频域经验似然推断程序，适用于平稳过程的谱密度中的参数分量。我们通过使用新的谱估计函数来构建经验似然函数，以实现对谱密度污染的鲁棒性。仿真研究证明了所提出的鲁棒频域经验似然方法的良好性能，该方法比普通经验似然对应物产生更准确的置信区域。