PENERAPAN METODE REGRESI KUANTIL PADA KASUS PELANGGARAN ASUMSI KENORMALAN SISAAN

Abstract

ABSTRACT

Quantile regression is a regression method with the approach of separating or dividing the data into any particular quantiles. The method is by minimizing the absolute weighted residual asymmetric and estimate the quantile functions conditional on distribution of data. Quantile regression parameter estimation does not require parametric assumptions. This study aims to apply the quantile method for data that violates the assumption of residual normality. Small size of data were generated from various distribution which residual designed with chi squared distribution. This study resulted manyproposed models that divided over several quantiles selected. The values of regression coefficient estimation were close to the initial value. This study found that the proposed model was good enough and could be accepted.

Keywords: Quantile Regression, Residual Normality, Regression Coefficient