Lossy Text Compression Techniques

Abstract

Most text documents contain a large amount of redundancy. Data compression can be used to minimize this redundancy and increase transmission efficiency or save storage space. Several text compression algorithms have been introduced for lossless text compression used in critical application areas. For non-critical applications, we could use lossy text compression to improve compression efficiency. In this paper, we propose three different source models for character-based lossy text compression: Dropped Vowels (DOV), Letter Mapping (LMP), and Replacement of Characters (ROC). The working principles and transformation methods associated with these methods are presented. Compression ratios obtained are included and compared. Comparisons of performance with those of the Huffman Coding and Arithmetic Coding algorithm are also made. Finally, some ideas for further improving the performance already obtained are proposed.