Mathematical modeling on the base of functions density of normal distribution

Abstract

One of the urgent tasks in many modern scientific studies is the comparative analysis of indicators that characterize large sets of similar objects located in different regions. Given the significant differences between the regions compared, this analysis should be carried out using relative indicators. The objective of the study was to use the density functions of the normal distribution to model empirical data that describe the compared sets of objects located in different regions. The methodological approach was based on the Chebyshev and Lyapunov theorems. The research results focus on the main stages of the construction of normal distribution functions and the corresponding histograms, as well as the determination of the parameters of these functions. The work possesses a degree of originality, since it provides answers to questions such as the justification of the necessary information base; performing computational experiments and developing alternative options for the generation of normal distribution density functions; comprehensive evaluation of the quality of the functions obtained through three statistical tests: Pearson, Kolmogorov-Smirnov, Shapiro-Wilk; identification of patterns that characterize the distribution of indicators of the sets of objects considered. Examples of empirical data models are given using distribution functions to estimate the share of innovative firms in the total number of firms in the regions of Russia.