Klasifikace s využitím převodu vektorových dat na sítě

Abstract

{This Master's thesis is focused on the classification of vector data using networks. The work provides an overview of the topics associated with the classification of vector data using networks and is primarily focused on the impact of the created networks on the result of the classification. The main goal of this work is to create an application that will allow to convert vector data to network, it classification and then the subsequent evaluation of classification results using an interactive network model. The experiments performed in this work will show that sparse networks are especially suitable for classification. Next, several network construction algorithms will be presented together with the networks they generate and then, subsequently, they will be compared. Finally, experiments will be performed to show that the number of vertices used fo classification that are necessary for good results can be calculated based on the properties of the network.