Comparison of Curve Estimation of the Smoothing Spline Nonparametric Function Path Based on PLS and PWLS In Various Levels of Heteroscedasticity

9th Annual Basic Science International Conference 2019

自然科学（总论）

Rinaldo Fernandes, Adji Achmad^1^5 ; Hutahayan, Benny^2 ; Solimun^1^5 ; Arisoesilaningsih, Endang^3 ; Yanti, Indah^4^5 ; Astuti, Ani Budi^1^5 ; Nurjannah^1^5 ; Amaliana, Luthfatul^1^5

Department of Statistics, Faculty of Mathematics and Natural Sciences, University of Brawijaya, Malang, Indonesia^1

Department of Bussiness Administrative, Faculty of Administrative Sciences, University of Brawijaya, Malang, Indonesia^2

Department of Biology, Faculty of Mathematics and Natural Sciences, University of Brawijaya, Malang, Indonesia^3

Department of Mathematics, Faculty of Mathematics and Natural Sciences, University of Brawijaya, Malang, Indonesia^4

Kelompok Kajian Unggulan Pemodelan Statistika di Bidang Manajemen, Faculty of Mathematics and Natural Sciences, University of Brawijaya, Malang, Indonesia^5

关键词: Curve estimation;    Heteroscedasticity;    Non-parametric;    Nonparametric approaches;    Nonparametric functions;    Relative efficiency;    Research results;    Smoothing spline;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/546/5/052024/pdf
DOI  :  10.1088/1757-899X/546/5/052024

学科分类：自然科学（综合）

来源: IOP

【 摘 要 】

Linearity assumption that has not been fulfilled in the path analysis should use nonparametric approach. This research uses smoothing spline nonparametric path analysis with generated data where the condition of heteroscedasticity level measured through MAPD statistic will be applied to the data. The conditions are MAPD 0.01 - 0.20; 0.21 - 0.40; 0.41 - 0.60; 0.61 - 0.80; and 0.81 - 1.00. The purpose of this research is to determine the comparison of curve estimation of spline smoothing nonparametric path function on every level of heteroscedasticity category (DM) and without considering the heteroscedasticity (TM). The research results found that relative efficiency value of DM (PWLS) estimator with TM (PLS) that is always more than 1 for every heteroscedasticity level and every observation size. Thus, it was obtained better DM estimator (PWLS approach) comparing to TM (PLS).