| Meteorological volumetric radar data are used to detect thunderstorms. The classification of storm cells is a difficult problem due to the complex evolution of them, the high dimensionality of the weather data, and the imprecision and incompleteness of the data. This thesis investigates the classification theory and approaches of rough set, and use them to classify different types of storm events. The rough set classification strategies are compared with other ones to determine which approaches will best classify the volumetric storm cell data coming from the Radar Decision Support System database of Environment Canada. The criterion for comparison is the accuracy coefficient in the classification over a testing data. The results obtained with the rough set approach show that they are a little better than other ones, in terms of accuracy, for the volumetric storm cell classification. |
