This paper considers using asymmetric kernels in local linear smoothing to estimate a regression curve with bounded support. The asymmetric kernels are either beta kernels if the curve has a compact support or gamma kernels if the curve is bounded from one end only. While possessing the standard...