Shlukování atributů

Abstract

This thesis is devoted to the topic of attribute clustering, which means in other words selection of the most relevant attributes from the original dataset. Attribute clustering is can be an important preprocessing step on solving classification on big data, which consists of up to hundreds to thousands of attributes. In this thesis, a few approaches for attribute clustering are described, especially algorithm Most Neighbors First. Also part of this thesis is the implementation of the application for attribute clustering and other related data analysis.