"Compared some of the semi-parametric methods in analysis of single index model


As the process of estimate for model and variable selection significant is a crucial process in the semi-parametric modeling At the beginning of the modeling process often At there are many explanatory variables to Avoid the loss of any explanatory elements may be important as a result , the selection of significant variables become necessary , so the process of variable selection is not intended to simplifying model complexity explanation , and also predicting. In this research was to use some of the semi-parametric methods (LASSO-MAVE , MAVE and The proposal method (Adaptive LASSO-MAVE) for variable selection and estimate semi-parametric single index model (SSIM) at the same time . The result that the best method for estimating and the variable selection of semi parametric single index model is proposal method (Adaptive LASSO-MAVE) of first model and (LASSO-MAVE) of second method useful for average mean squares error (AMSE).