CUBIC SPLINE ESTIMATOR IN MULTIPREDICTOR NONPARAMETRIC REGRESSION MODELS WITH LOGNORMAL ERRORS AND ITS

Nur Chamidah,
Toha Saifudin,

Matematika dan Sains
Vol 16, No 1 (2009)

Suppose that n observations follow multiplicative nonparametric regression models with errors which are lognormally distributed. The assumption on causes the values of ln would be normally distributed . So, by taking natural logarithm of the model, we have an additive nonparametric regression model. In this paper, we estimate regression function of the model by using nonparametric regression approach, i.e, cubic spline estimator. Next, we give an applying illustration of the model on Gmelina Arborea Roxb data.