Model of Human Development Index in West Nusa Tenggara Province using Geographically Weighted Ridge Regression Method

Abstract

The Human Development Index (HDI) is a comparative measurement of life expectancy, education, and living standards for all countries that aim to classify a country as a developed, developing, or underdeveloped country. The Regency/City Human Development Index (HDI) value in West Nusa Tenggara Province has increased in the last five years. It is necessary to analyze the factors that cause the increase in the value of HDI. This study aims to determine the HDI model with six variables that affect HDI. HDI data is included in spatial information obtained from several regions. This study used ten observation areas: West Lombok Regency, Mataram City, Central Lombok Regency, North Lombok Regency, East Lombok Regency, Sumbawa Regency, West Sumbawa Regency, Dompu Regency, Bima Regency, and Bima City. The Geographically Weighted Ridge Regression (GWRR) method was used in modeling the HDI value. This method can be used for data that have multicollinearity problems. The striking difference between the GWRR method and the multiple linear regression method is the presence of weights. Therefore, the Weighted Least Square (WLS) method was the parameter estimation method used, with the weighted fixed Gaussian. As a result of the differences in the observation area, the models produced in this study consist of ten different models. The modeling was done by considering the Mean Square Error (MSE) values and Akaike's Information Criterion (AIC). The results of the analysis carried out show that the variables that have multicollinearity problems are the variables X1, X4, X5, and X6. The GWRR model obtained gives the MSE value of the GWRR method of 0.158861 and the AIC value of the GWRR method of -87.36997, with variables that significantly affect the value of the HDI data, are the expenditure variable per capita and the average length of schooling.