Převod vektorových dat na data síťová

Abstract

This master's thesis describes network construction from vector data and methods which are used to construct network data from vector data in a few steps. This conversion is used in Semi-Supervised Learning area when we have dataset in which some data are labeled and the rest are unlabeled. Conversion to network allows classify the unlabeled data. The main goal of this thesis is implementation of methods for network construction, experiments on datasets and comparsion of methods for network construction.