Multivariate Local Polynomial Kernel Estimators: Leading Bias and Asymptotic Distribution
在更平滑的回归函数假设下，推导了多元局部多项式核估计量整个向量的显式主导偏差项，可用于优化条件均值及其导数的平滑参数。
Masry (1996b) provides estimation bias and variance expression for a general local polynomial kernel estimator in a general multivariate regression framework. Under smoother conditions on the unknown regression function and by including more refined approximation terms than that in Masry (1996b Masry , E. ( 1996b ). Multivariate local polynomial regression for time series: uniform strong consistency and rates . Journal of Time Series Analysis 17 : 571 – 599 .[Crossref] , [Google Scholar]), we extend the result of Masry (1996b Masry , E. ( 1996b ). Multivariate local polynomial regression for time series: uniform strong consistency and rates . Journal of Time Series Analysis 17 : 571 – 599 .[Crossref] , [Google Scholar]) to obtain explicit leading bias terms for the whole vector of the local polynomial estimator. Specifically, we derive the leading bias and leading variance terms of nonparametric local polynomial kernel estimator in a general nonparametric multivariate regression model framework. The results can be used to obtain optimal smoothing parameters in local polynomial estimation of the unknown conditional mean function and its derivative functions.