Summary
This Diploma Theses summarizes and describes the classification methods of the quantitative data in the geographical information systems that are available in present days.
The accessibility of the methods was proved in six GIS programs. Three of them were commercial (ArcGIS 9.3, Geomedia Professional 6.0 and AutoCAD Map 3D 2009) and the rest of them was open sources (GRASS 6.4.0, Quantum GIS 1.0.1 and uDig 1.1.0-SC3).
All of these programs were compared with regard to the classificaton methods. The main criterion was the number of the implemented methods. Software ArcGIS 9.3 can work with the highest number of methods – with eight methods. The second place occupies AutoCAD Map 3D and GRASS that can work with six methods. It means that GRASS is the best open source software for the data classification of all tested programs.
The suitability of the methods use was tested for two sets of data with logarithmic and normal distribution. The method was visual evaluation of classifiation methods. This doesn´t warrant the objective results, but it is sufficiently for the suitable methods setting.
As the suitable classification methods for logaritmic division were set: the quantile method ( in this case the quintil method) and the geometrical classification method.
The quantile method and the Armstrong-Xiaovo criterion (Natural Breaks Jenks) were the best for the chosen data in the normal division. The geometrical classification method provided the satisfactory results. The results show that it is better to use more than one suitable method for the data in normal division. The method of the equal intervals was unsuitable for both types of the data. This method would be suitable for the data in the constant distribution.
The new classification method was implemented to the ArcGIS 9.3. The script was written in the MS Visual Basic 6.3. The script offers not only the new methods, but also the methods which are already integrated.
The script with the new classification method was inserted to the toolbar. The whole script is commented in english. The toolbar and the source code will be able to download on the website of the diploma thesis, on the CD-ROM in Department of geoinformatics in Palackého University in Olomouc.