MODEL REGRESI NONPARAMETRIK ADITIF DENGAN FUNGSI LINK

Abstract



If  given a couple of data  and  the  relationship  between  variable x and y is revealed as nonparametric regression model with is an error which is random variable with zero mean and variance  and is unknown function. If is an additive function with x  R^d, d ≥ 2. By using link function F, this research is focused on the way to get the estimator in additive component with link function by using two-stage methode and completed with simulation using logit link function. The result shows that two-stage estimator is eisier be used to model nonparametric regression by link function rather than Nadaraya-Watson and local linear. From the simulation, it’s obtained that of the mean and R² value of the estimator is similar.