Local Polynomial Regression is one of the important methods in estimation nonparametric regression curve, furtherwere containing outliers in the data sample will affect the estimated parameters and will also affect the shape of the estimated curve. Therfore, many robust methods have been appeared which will belittle the influence of outliers by down the weighting given for outliers in data samples. One of these methods is LOWESS method. In this paper we studied the influence of degree of polynomial in finding estimators using LOWESS method. Depending on the generated data through simulation study and by using one of error criterion which is integrated squared error (ISE), we found out quadratic polynomial was better than constant and linear polynomial when estimating regression curve using LOWESS method.
Nonparametric Regression, Local Regression, Robust, Outliers, LOWESS.
Cite this paper
Hafed M. Motair. (2020) On the Degree of Polynomial for Robust Locally Weighted Polynomial Regression. International Journal of Mathematical and Computational Methods, 5, 26-37
Copyright © 2020 Author(s) retain the copyright of this article.
This article is published under the terms of the Creative Commons Attribution License 4.0