Geographically Weighted Regression Model with Covariate Measurement Errors

  • Ida Mariati Hutabarat Department of Mathematics, Faculty of Mathematics and Natural Sciences, Cenderawasih University, Jayapura, Indonesia
  • Yacob Ruru Department of Mathematics, Faculty of Mathematics and Natural Sciences, Cenderawasih University, Jayapura, Indonesia
Keywords: Mixed linear model, restricted maximum likelihood method, asymptotic normality properties, weighting matrix

Abstract

In this paper, we determine the parameter estimators of the Geographically Weighted Regression (GWR) model with measurement errors through the mixed linear model approach. In contrast to model that does not pay attention to geographic location factors, GWR is very concerned about geographic location factors. This model will produce the parameter estimators of the local model for each point or location where data is collected. The mixed linear model approach in the GWR model is form a weighting matrix for each observation location. The estimation method used to estimate parameters in the mixed linear model is Restricted Maximum Likelihood (REML) method. Asymptotic normality properties of the estimators are obtained. The estimators are shown to be consistent. From the estimation results obtained, we illustrate the data of malnutrition sufferers in East Java province.

Downloads

Download data is not yet available.
Published
2021-03-29
Section
Articles