UJI HIPOTESIS MODEL MIXED GEOGRAPHICALLY WEIGHTED REGRESSION DENGAN METODE BOOTSTRAP

Abstract

One of the models that can be used to analyze the spatial data is the Mixed Geographically Weighted Regression (MGWR) model. MGWR was used to solve the problem where certain predictor variables are influencing the response globally while others are locally. This paper tried to estimate the parameters of MGWR model using Weighted Least Square method, and then get the p-value of statistical test using the bootstrap procedure. The application of the bootstrap methode in this case is resample the statistical test base on the residual model. Key words: Bootstrap procedure, Mixed Geographically Weighted Regression, p-value, Weighted Least Square