Arabic Text Compression Technique Based on Lempel-Ziv Algorithm

摘要

This paper presents a new Arabic text compression technique based on Lempel-Ziv algorithms. The new technique preprocesses an input file to produce binary data streams and then submits the streams to a bitwise compression algorithm. Preprocessing is performed by constructing a character mapping table in which each character in the text file is mapped into a weighted ASCII-like code. The table is then used to convert the input file into a binary equivalent which is then condensed on the bit level. The proposed technique was examined using a popular Arabic text file. The results showed that using extension-order of 8 or 4, the new technique achieved a compression ratio within the range of 4.25 to 4.7 bits/character, respectively. The improvement in compression efficiency is due to the significant reduction in the number of symbols in the generated file.