Využití spojitých rozdělení jako modelů mzdových rozdělení

This thesis deals with the description of continuous probability distributions suitable for modeling empirical wage distributions. In the theoretical part, the reader gets acquainted with the theoretical two-parameter, three-parameter and four-parameter continuous probability distributions and their main characteristics of location, variability and shape. The second part deals with the theoretical description of the maximum likelihood estimation method and the minimum distance estimation method of parameters of selected continuous distributions. The last part of the work is focused on the practical application of more detailed probability distributions and methods of point estimation of parameters to data from structural surveys of wages of employees for 2020. Nine probability distributions were examined, their parameters were estimated by two methods, so there were a total of eighteen models. The models were compared using the value of the likelihood function, the Akaike information criterion, and the sum of the absolute or squared deviations of the observed and theoretical probabilities for all gross monthly wage intervals. Models with more parameters proved to be mostly more accurate, but this was definitely not always the case.