| Abstract [eng] |
The main problem in this paper is to find the estimations of the rate of not equal convergence in the maximum and minimum schemes using the Burr distribution. The reason why this distribution was chosen is because of the relationships between the Burr distribution and various other distributions, like the log-logistic distribution, the Pareto distribution. This is very important, because the Burr distribution is the main one of them. The extreme value theory is used to find the limit theorems. It says that if there are a lot of components in the structure of data, then to simplify the modeling of data we can change the Burr distribution to the limit distribution. Then we can make the model which is based on the observations of maximum or minimum values of data. If the number of components in the structure of data is unknown then we should know how these components are distributed. So now we can approximate our distribution with the limit distribution by using the theorems of the limit distributions. In this case we have to compute the occurring errors. Then we are ready to find the estimates of errors. In this work errors are estimated theoretically. In conclusion we can state that our estimations of errors are of the same magnitude as 1/n. |
