PENERAPAN METODE EMPIRICAL BEST LINEAR UNBIASED PREDICTION (EBLUP) PADA MODEL PENDUGA AREA KECIL DALAM PENDUGAAN PENGELUARAN PER KAPITA DI KABUPATEN BREBES

Abstract

The coming of a policy about regional autonomy makes district government's choices of strategy and policy become crucial and important for it's district's development and prosperity. Indicator that can states this district development is Human Development Index (HDI). One of dimension that being used to predict the value of HDI is the dimensions of decent living, which can be shown from expenditure per capita. Should the samples of expenditure per capita are less than needed, it can cause difficulty to analyze the value of HDI on next level, which is sub-districal HDI. Direct estimaton only will not give enough validity for the results which can cause the increasing value for it's variance. Another method that can be used is small area estimation (SAE) with Empirical Best Linear Unbiased Prediction (EBLUP) method. This estimation uses the information from it's surrounding areas that correlates with the subject's parametrics. The evaluation for the results is done by comparing the value of Relative Root Mean Square Error (RRMSE) from a direct estimation with the RRMSE from an indirect estimation, which is the EBLUP method. Results from EBLUP estimation is better with average of RRMSE of 7,219% than direct estimation's average of RRMSE with 9,361%. Keyword : Expenditure per capita, Small Area Estimation (SAE), Empirical Best Linear Unbiased Prediction (EBLUP)