Abstract:
In this paper dealt with the study the method of choosing the best model of quantile regression through the use of penalty methods within the model of linear quantile regression . Where the researcher used two penalty methods, and these two methods are Lasso penalty function and Laplace error penalty function) This research dealt with the study of the method of choosing the best model of quantile regression through the use of penalty methods within the model of linear quantile regression. Where the researcher used two penalty methods, and these two methods are Lasso penalty function and Laplace error penalty function). We conducted one Monte Carlo simulation experiment with the assumption of the existence of a real vector for the parameters to be estimated, and this experiment was conducted with the assumption of generating different sample sizes in order to improve each method's level of accuracy. The results of the Monte Carlo simulation indicate that the Laplace penalty function method gave the lowest value to the average square of errors and the positive and negative parameters(FNR,FPR), and thus is better than the Lasso method .
|