Vyhledávání Vzorů v Komprimovaných Datech

Abstract

This Bachelor's thesis is about pattern matching. Main objective is to describe selected algorithms and data structures, that are used in practice for pattern matching on non-compressed as well as compressed data. Integral part of this thesis is subsequent implementation of the selected data structure. At present, compression algorithms using Burrows-Wheeler transformation are used extensively and data structure FM-Index depends on it. This data structure will be implemented in programming language C# and subjected to experiments. Experiments will mainly cover speed of pattern matching and will be cross examined against more classical algorithms. Space requirements will be tested on data of varying formats as well as with different configurations of FM-Index. At the end the results and findings from the experiments will be presented.