Kris De Brananter
Department of Statistics
Iowa State University

ABSTRACT:

Convergence Rates for Uniform Confidence Intervals Based on Local Polynomial Regression Estimators

We investigate the convergence rates of uniform bias corrected confidence intervals for a smooth curve using local polynomial regression for both the interior and boundary region.

We discuss the cases when the degree of the polynomial is odd and even. The uniform confidence intervals are based on the volume-of-tube formula modified for biased estimators.

We show that the proposed uniform confidence intervals attain, at least approximately, nominal coverage. Finally, we investigate the performance of the volume-of-tube

based confidence intervals for independent non Gaussian errors.

Refreshments at 3:45pm in Snedecor 2101.