Analysis of semi-parametric single-index models by using MAVE-method based on some kernel functions

Fayyadh Ali, Tareq Salih
2017 International Journal of Advanced Statistics and Probability  
In this paper, we used many forms of kernel functions with minimum average variance estimation (MAVE) method [Xia2002] we called the proposed methods (MAVE-Biweight), (MAVE-Epanechnikov ) and .(MAVE-Gaussian ) for estimation the parameters and the link function of the single – index model (SIM) comparing with other methods of estimation . to evaluate the performing of the various methods s simulation and a real data have been used, conclusions showed that the (MAVE- Gaussian) method in this
more » ... r gave better results compared with other methods depending on the mean squared error (MSE) and mean Absolute error (MAE) criterion for comparison.
doi:10.14419/ijasp.v5i1.7258 fatcat:2v3nlkmkhzaojolphi42nhl6ue