Parametric Quantile-Regression Models and Those Similar to Them

Abstract

The parametric approach to the construction of quantile regression models based on the selection and evaluation of parameters of the functional dependence of the moments of the first and second order distribution of the dependent variable from the values of the independent variable is considered. The mathematical and algorithmic maintenance of the procedures for the construction of parametric models is described. Comparative analysis of the statistical validation criteria distribution for normality is carried out. The information criteria that ensure selection of the optimal number of model parameters are given. The possibility of using the parametric quantile regression models for charting the age dynamics of the growth of children is considered.