Smoothing Parameter Selection Problem in Nonparametric Regression Based on Smoothing Spline: A Simulation Study
M. Seref Tuzemen
This study presented a comparative study of the different smoothing parameter selection methods. The parameter selection problem has been examined in respect to a smoothing spline implementation in predicting nonparametric regression models. For this purpose, a simulation experiment has been carried out by using a program written in MATLAB 6.5. The simulation experiment provides a comparison of the six different selection methods. In this context, 500 replications have been carried out in simulation for sample sets with different sizes. Thus, the empirical performances of the six selection criteria have been investigated and the suitable selection criteria are obtained for an optimum parameter selection.
Cited References Fulltext