Open AccessPemodelan Data Time Series dengan Pendekatan Regresi Nonparametrik B-Spline
Author(s) -
Zulaiha Rahasia,
Resmawan Resmawan,
Dewi Rahmawaty Isa
Publication year - 2020
Publication title -
aksioma: jurnal matematika dan pendidikan matematika/aksioma
Language(s) - English
Resource type - Journals
eISSN - 2579-7646
pISSN - 2086-2725
DOI - 10.26877/aks.v11i1.4903
Subject(s) - knot (papermaking) , nonparametric statistics , smoothing spline , spline (mechanical) , mathematics , nonparametric regression , statistics , computer science , econometrics , spline interpolation , engineering , structural engineering , chemical engineering , bilinear interpolation
Spline is one of the nonparametric approach, to adjust data so the final model has good flexibility. The purpose of this research is to model the time series data in the form of currency exchange rates by using the nonparametric B-spline approach. In B-spline modelling, determination of the order for the model, and the number and the placement of the knot are the criteria that must be considered. The best B-spline model obtained based on the selection of the optimal knot points with minimum Generalized Cross Validation (GCV) criteria. The modelling in this research use data on the exchange rate of the rupiah toward the US dollar in the period January 2014 - December 2018. The best B-spline model obtained by the 2 point knot approach, at points 11935.10 and 12438.29, with GCV valueequals to 55683.09.Keywords: Nonparametric Regression; B-Spline; Generalized Cross Validation