ESTIMATOR POLINOMIAL LOKAL UNTUK PEMODELAN HARGA EMAS DI INDONESIA MENGGUNAKAN REGRESI NONPARAMETRIK

No Thumbnail Available

Date

2022-05

Journal Title

Journal ISSN

Volume Title

Publisher

Elfitra

Abstract

This paper discusses about the nonparametric regression model based on local polynomial estimators on gold prices in Indonesia. Local polynomial estimators can be obtained by minimizing Weighted Least Square (WLS). Optimal bandwidth can be obtained using the Generalized Cross Validation (GCV) method. Furthermore this methods are applied to nonparametric regression models based on local polynomial estimators in the case of gold prices in Indonesia. Gold price modeling in Indonesia using local polynomial regression with Gaussian kernel function obtained optimal bandwidth of 10. The model produces a minimum GCV score of 854882843 with the optimum order of the local polynomial is order 4. The goodness of fit model used is the Mean Absolute Percentage Error (MAPE) with MAPE 6.182%, that is smaller them 10%. Then local polynomial regression model order 4 for gold price is very accurate.

Description

Keywords

Nonparametric regression, local polynomial estimator, Weighted Least Square, Generalized Cross Validation, Mean Absolute Percentage Error

Citation

Perpustakaan

Collections