search for




 

Nonparametric estimation for motorcycle data in scale-space
Journal of the Korean Data & Information Science Society 2020;31:109-21
Published online January 31, 2020;  https://doi.org/10.7465/jkdi.2020.31.1.109
© 2020 Korean Data and Information Science Society.

Jib Huh1

1Department of Statistics, Duksung Women’s University
Correspondence to: Professor, Department of Statistics, DuksungWomen’s University, Seoul 01369, Korea. E-mail: jhuh@duksung.ac.kr
This research was supported by the Duksung Women’s University Research Grants 2018.
Received December 27, 2019; Revised January 8, 2020; Accepted January 8, 2020.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
The bandwidth of kernel type estimators, which is the smoothing parameter, is an important factor in determining the smoothness of the estimated function. Chudhuri and Marron (1999) proposed SiZer (SIgnificant ZERo crossings of derivatives) tool based on visualizing the testing results as a color map to determine significance of increasing or decreasing trend of function using a wide range of bandwidths rather than selection of bandwidth. In this paper, SiZers for the regression and log-variance function are described and the performance of SiZer for the motorcycle data using the package SiZer in R are visually demonstrated.
Keywords : Bandwidth, confidence interval, Gaussian kernel, local polynomial estimator, SiZer.