An algorithm for determining the parameters of the distribution density function with the application of nonparametric statistics methods.

The paper considers an algorithm for determining the unknown function of random variable distribution density presented in the form of a finite trigonometric series. The random variable is given by a sample of values. To calculate the coefficients of the series, we use a set of quantile values of a random variable, obtained on the basis of the function of random value distribution density, recovered from the sample by means of Parzen-Rosenblatt estimates. The problem is reduced to solving a system of equations that are linear towards the desired coefficients. The developed algorithm is exemplified by experimental data processing. [ABSTRACT FROM AUTHOR]