Empirical likelihood confidence intervals for local linear smoothers

Empirical likelihood is considered in conjunction with the local linear smo other to construct confidence intervals for a nonparametric regression func tion with bounded support. The coverage error of the empirical likelihood c onfidence intervals is evaluated and is shown to be of the same order throu ghout the support of the regression function. This is a significant improve ment over confidence intervals based directly on the asymptotic normal dist ribution of the local linear estimator, which have a larger order of covera ge error near the boundary. This improvement is attributable to the natural variance estimator that empirical likelihood implicitly chooses for the lo cal linear smoother.