Yazar "Ozturk, Ibrahim" seçeneğine göre listele

Listeleniyor 1 - 1 / 1
  • [ X ]
    Öğe
    A Comparison of Text Compression Performance of Statistical Coding Methods in Turkish and English
    (Kırıkkale Üniversitesi, 2023) Ozturk, Ibrahim; Kaya, Hakan Celil
    Data compression is the set of operations performed to enable digital data to occupy less space than it does in memory. These operations are carried out by utilizing more or less repeated data chunks depending on the file type. In this way, compression operations allow for a more efficient use of the memory and data communication bus. Compression techniques are divided into two groups: lossless and lossy compression. Lossless compression includes dictionary-based coding and statistical coding methods. Statistical coding represents the most frequently occurring characters in the data with shorter codewords, while less common characters are represented by longer codewords. Although the frequency of character use is at the heart of statistical coding methods, the processing steps differ depending on the method used. In this study, the performances of the Huffman, Shannon-Fano, and Arithmetic coding methods, which use statistical coding for compression, were compared on English and Turkish texts. Text-based files within the Calgary corpus for English and compilations of newspaper columns for Turkish are used in the study. The comparisons are made based on savings rate, compression-decompression times, Bit per character (BPC), and entropy metrics. The results show performance differences in savings rate, BPC, and entropy metrics between statistical coding methods for English and Turkish texts.