Modeling Poverty Percentages in the Papua Islands using Fourier Series in Nonparametric Regression Multivariable

Abstract

Nonparametric regression has high flexibility in estimating the regression curve. Estimation techniques that are quite popular in nonparametric regression are Fourier Series estimators. Fourier series are trigonometric polynomials that have flexibility, so they can adapt effectively to the local nature of data. This Fourier series estimator is generally used if the data investigated by the pattern is unknown and there is a tendency for repetitive patterns to follow the trend line. The purpose of this study is to examine Fourier Series estimates and apply to cases of poverty in the Papua Islands. The Papua Islands which consist of 2 Provinces namely Papua Province and West Papua Province. Papua Province ranked first and West Papua Province ranked second in poverty in Indonesia. Based on nonparametric regression multivariable modeling using Fourier series, this model is good. This can be seen from the value obtained by Generalized Cross Validation (GCV) = 80, 76, Mean Square Error (MSE) is 18, 31 and R2=78, 16%.