Kernel estimation of the regression mode for fixed design model

In this paper, we study the problem of estimating nonparametrically the regression mode for fixed design model. We suppose the error random variables are independent. The joint asymptotic normality of the regression mode estimator at different fixed design points is established under some regularity conditions. The performance of the proposed estimator is tested via a simulation study.

本文研究固定设计模型（fixed design model）下的非参数回归众数（regression mode）估计问题。我们假设误差随机变量相互独立。在若干正则性条件下，建立了不同固定设计点处的回归众数估计量的联合渐近正态性。通过模拟实验对所提出的估计量的性能进行了验证。