Comparison of Selected Text Compression Algorithms

U ovom radu opisani su i uspoređeni odabrani algoritmi za kompresiju. Algoritmi su podijeljeni u tri kategorije: entropijski algoritmi, algoritmi temeljeni na metodama rječnika i napredni algoritmi. Prvu kategoriju predstavljaju Huffmanovo kodiranje i Shannon-Fanov algoritam. Sljedeću kategoriju predstavljaju LZW i LZSS algoritmi. Posljednju kategoriju predstavljaju DEFLATE i LZMA algoritmi. Svi algoritmi implementirani su u dva vrlo popularna programska jezika – C++-u i Pythonu. Uspoređivani su omjeri kompresije koje algoritmi ostvaruju na odabranim ulaznim tekstovima. Osim toga, uspoređivana su i vremena izvršavanja algoritama s obzirom na jezik implementacije. Rezultati su prikazani tablično i grafički. Na kraju su navedeni zaključci koji su izvedeni iz rezultata.In this final paper, selected compression algorithms are described and compared. Algorithms are divided into three categories: entropy algorithms, algorithms based on dictionary methods, and advanced algorithms. The first category is represented by Huffman coding and Shannon-Fano algorithm. The next category is represented by LZW and LZSS algorithms. The last category is represented by DEFLATE and LZMA algorithms. All algorithms are implemented in two very popular programming languages – C++ and Python. The compression ratios achieved by the algorithms on the selected input texts were compared. In addition, the execution times of the algorithms, with respect to the programming language, were compared. The results are presented tabularly and graphically. At the end, the conclusions based on the results are stated