Journal of Applied Mathematics & Bioinformatics

Robust Parameter Estimation Methods: L_Moments and TL-Moments of Probability Distributions

• [ Download ]
• Times downloaded: 10146

• Abstract

   

  This paper deals with the application of such robust methods of point parameter estimation, as the methods of L-moments and TL-moments on economic data. The advantages of these highly robust parametric estimation methods are aware when applied to small data sets, especially in the field of hydrology, meteorology and climatology, in particular considering extreme precipitation. The main aim of this contribution is to use these methods on large datasets, and comparison the accuracy of these two methods of parametric estimation with the accuracy of the method of maximum likelihood, especially in terms of efficiency of parametric estimation. The study is divided into a theoretical part, in which mathematical and statistical aspects are described, and an analytical part, during which the results of the use of three robust parametric estimation methods are presented. Total 168 income distributions of the years from 1992 to 2007 in the Czech Republic (distribution of net annual income per capita in CZK) were analyzed. There are a total income distribution for all households of the Czech Republic together and further the income distributions broken down by gender, job classification, classification of economic activities, age and educational attainment. Three-parametric lognormal curves represent the basic theoretical probability distribution.