Predictive Text Compression with Self-Organizing List
スポンサーリンク
概要
- 論文の詳細を見る
The paradigm of data compression using Markov model or prediction technique has shown its capability of high compression rates.Such a paradigm separates compression model into two units;modelling and encoding.The modelling part provides a prediction in the form of probability distributions to the encoding part producing a shorter representation.There is a number of algorithms using the prediction techniques,e.g.,PPM[2],DAFC[5],UMC[6],DHPC[7],DMC[3],and etc.All of mentioned algorithms used prediction techniques to bias the probabilities and employed arithmetic coding.Recently we have proposed the BSTW with PPM algorithm[8]which applies the prediction technique to BSTW[1].The algorithm reduces unit of "word" used in original BSTW to"character"and applies prediction techniques to reorganize code tables.The characters with greater probabilities will be placed nearer to the beginning of code table than characters with smaller probabilities.The coding scheme will take the.advantage of shorter encoding for characters near the beginning of the list.Although the algorithm shows fairy good performance in compression ratio,the problems of processing time and memory space still remain unsatisfactorily resolved.In this presentation,to speed up the processing time,the predictive data compression with self-organizing list,e.g.,move-to-front,transpose,moveahead-k will be introduced.The expenmental results are also presented to demonstrate their performances.
- 一般社団法人情報処理学会の論文
- 1992-09-28
著者
関連論文
- Predictive Text Compression with Self-Organizing List
- The Zero-Frequency Problem in Predictive Self-Organizing Lists Data Compression
- BSTW Data Compression with PPM