Perbandingan Penggunaan Metode Analisis Regresi Ridge dan Metode Analisis Regresi Komponen Utama dalam Menyelesaikan Masalah Multikolinieritas (Studi Kasus Data PDRB Propinsi Sumatera Utara)

Abstract

Multicollinearity is a condition where there is a regression in a very high correlation between the independent variables. Ridge Regression Analysis and Principal Component Regression analysis is a method to solve the multicollinearity that occurs in multiple regression analysis. Ridge Regression Analysis is a method that gives a relatively small constant bias by multiplying the constant bias on the diagonal identity matrix θ, so the estimation parameter be: . Principal Component Regression analysis is basically aimed to simplify the variables observed by shrinking (reduced) the dimension. This is done by removing the correlation between independent variables through the transformation of the independent variables of origin to a new variable that does not correlate at all, or so-called principal component (principal component). Testing coefficients obtained from the two methods would indicate that multicollinearity in a multiple linear regression was completed.