REGRESI NONPARAMETRIK KERNEL ADJUSTED

Abstract

Nadaraya Watson's kernel adjusted regression estimator is an estimator whose kernel is taken from the family of scale-location associated with the classical kernel density estimator. Based on these estimator, it can be obtained optimal bandwith and scale parameter. This estimator gives a better estimation results compared with Naradaya Watson's classical kernel regression estimator. This is proven by the small grade MSE which is given by this estimator.