Single and Binary Performance Comparison of Data Compression Algorithms for Text Files

Abstract

Data compression is a technique used to reduce the size of a file. To reduce the size of a file, unnecessary information is removed or parts that repeat the same information are stored once. When the compressed file is opened, it has all the properties of the original file and can be used in the same way. Data compression can be done using different techniques. Some of these techniques include Huffman coding, Lempel-Ziv-Welch (LZW) coding and BWT (Burrows-Wheeler Transform). Some techniques simplify the data, while others identify repetitive parts of the data and store them in a separate file. Which technique to use depends on the type and size of the data to be compressed. Huffman, LZW, BWT and deflate algorithms are the most commonly used techniques for text compression. Each algorithm uses different approaches and can produce different results in terms of compression ratios and performance. In this study, different data compression techniques were measured on specific datasets by using them individually and in stacked pairs. The most successful result was obtained with the deflate algorithm and 29.08 compression ratio when used individually. When considered in the form of stacked binaries, the BWT-Deflate binary gave the best result with a compression ratio of 57.36. In addition, when compressing in pairs, which algorithm is applied first and which one is applied later can make a significant difference in the compression ratio. In this study, the performance measurements obtained by applying the algorithms in different orders are compared and suggestions are presented to achieve optimum performance.