Comparison of Newton Raphson Method and Ridge Method In Probit Regression Parameter Estimation

Abstract

Probit regression model is a non-linear model used in the process of analyzing the relationship between a response variable that has categorical properties. The problem that is very often experienced in probit regression when the predictor variable consists of one or more is that there is a very high correlation between predictor variables called multicollinearity. To overcome this, the Newton Raphson method and the Rigde method are used. So this research was conducted to compare the Newton Raphson method and the Ridge method in the estimation of the Probit Regression parameter. The data used in this research is 1000 data generation that contains multicollinearity. Based on this research, the estimated mean square error of the Probit Regression model using the Newton Raphson method is 0.488. The estimation result of the mean square error of the Probit Regression model using the Ridge method is 0.488. The results of this study indicate that the estimation of the Probit Regression parameter using the Newton Raphson method is as good as the Ridge method. This can be seen from the estimated value of MSE using the Newton Raphson method and the Ridge method. This can happen due to the small value of the langrage multiplier obtained, so it does not have an impact on the model obtained.