Використання методу k-середніх для ідентифікації моделей неоднорідних розподілів випадкових величин

Abstract

In the article the problem of identification of inhomogeneous distribution models of random variables is considered. The procedure, which provides a preliminary estimation of the quantity and parameters of components and their further improvement with the help of methods of minimization of the Kolmogorov−Smirnov test or the omega-squared test calculated values, is proposed. The results of analysis of empirical data about the elections of the President of Ukraine in 2010 year.