Vít VOŽENÍLEK, Jaromír KAŇOK, Pavel TUČEK

Detekce, prokazatelnost a vizualizace extrémů dat ve statistických souborech

Voženílek, V., Kaňok, J., Tuček, P.: Extreme detection, provableness and visualization of statistical data files. Kartografické listy, 2008, 16, 9 figs., 17 refs.

Abstract: The aim of the article is to make one known, that the extreme values of the observed geographical variable could be visualized in many ways. Authors introduce the optimal procedures for visualization of the extreme values in demographic datasets. The methodology is based on the statistical preprocessing of the data. The detection of the extreme values must be performed as first. The visualization of these values could be drawn after the significant prove. There is presented boundary which could be computed from the dataset and could serve as a signal point for the extreme values in the datasets. The extreme values are divided into two groups according to the point of occurrence. The first group is reserved for the extreme values in the frequency area. The second group describes the extreme values in the data area. The conclusion is dedicated to the measuring of the entropy in order to obtain the significant number of intervals for the creating of the maps with the respect to the occurrence of the extreme values.

Keywords: map, errors, outliers, extremes, visualization