Induction of user preferences in semantic web

This thesis deals with the induction of user preferences. In its first part it surveys the field of user preferences on semantic web. Subsequently, it describes various methods of data collecting. In the part concerned with inductive methods the paper focuses mainly on collaborative filtering and decision trees. Many value models and preference relations belong among the most important of the described models. The aim of the thesis is neither to provide a complete enumeration in its individual parts, nor to render as detailed view as possible. Rather, it aims at gaining a basic outline of this broad theme. The second part of the thesis pursues the description of the implementation of a testing framework which enables the easiest testing of inductive methods without the knowledge of the inner structure of the framework. The final part describes the technique and results of the testing of decision trees and collaborative filtering on databases NetFlix and MovieLens. Here the framework described in the second part of the paper has been used as a testing tool