METODE PRINCIPAL COMPONENT ANALYSIS UNTUK MENGATASI MULTIKOLINEARITAS PADA REGRESI LINIER BERGA NDA (STUDI KASUS FAKTOR YANG MEMPENGARUHI INDEK PEMBANGUNAN MANUSIA DI JAWA TIMUR)

Abstract

This research is conducted in order to implement PCA (Principal component Analysis) to overcome multicollinearity in multiple linear regression. The case studies used are factors that affect the human development index (HDI) in East Java by 2013. These factors consist of the old school, literacy rates, spending per capita and life expectancy, so there are four (4) independent variables in this research. This research is using data publication because it is non-reactive research that is the kind of research is devoted to the publication of data. The result is obtained equation regression model Y = 72,440 + 4,470 F1, with new independent variable F1= 0,292 (Old School) + 0,286 (Literacy Rates) + 0,267 (Spending Per Capita ) + 0,249 ( Life Expectancy).

The conclusion of this research is regression results obtained from the new independent variable (F1) with the human development index variable (Y) indicate a strong influence because of the significant value p= 0.000 <0.05. In addition the results of the analysis of PCA (Principal Component Analysis) can also be used to determine the dominant factor. The dominant factor of human development index case in East Java by 2013 that is old school with coefficient 0.292.

Keywords: multicolinearity; human development index; PCA (Principal component Analysis)