Comparison of Estimates Nonparametric In Multiple Regression Analysis Function (Gamma ,Beta)


Abstract:- The use of non-parametric models and subsequent estimation methods requires that many of the initial conditions that must be met to represent those models of society under study are appropriate, prompting researchers to look for more flexible models, which are represented by non-parametric models In this study, the most important and most widespread estimations of the estimation of the nonlinear regression function were investigated using Nadaraya-Watson and Regression Local Ploynomial, which are one of the types of non-linear capabilities and using Gamma Kernel, Beta Kernel functions compared with the Monti-Carlo simulation method Different variations and sizes of different samples. The simulation results using the Monti-Carlo method showed that the best estimate was Nadaraya-Watson and for all cases.