Efektivní implementace komprese blokovým tříděním

Abstract

This Master thesis deals with an effective implementation of block-sorting compression. The most important part of this compression is the Burrows-Wheeler transformation, which transforms input string into more suitable sequence. The main step of this transformation is sorting all rotations of the original data. This is the most time and memory consuming procedure, so the effectivity of an algorithm depends mainly on this operation. Early approaches were basically based on standard comparison function, later on started to appear more sophisticated algorithms based on suffix trees and suffix arrays.