Yu, C.-L.C.-L.YuJA-LING WU2020-06-162020-06-161998https://www.scopus.com/inward/record.uri?eid=2-s2.0-0032155169&doi=10.1016%2fS0165-1684%2898%2900096-6&partnerID=40&md5=0736987866f0055cb93953ca655ca14bIn this paper, an adaptive multi-dictionary model for data compression is proposed. Dictionary techniques applied in lossless compression coding can be modeled from the dictionary management point of view which is similar to that of cache memory. The behavior of a compression technique can be described by nine parameters defined in the proposed model, which provides a unified framework to describe the behavior of lossless compression techniques including existing probability-based Huffman coding and arithmetic coding, and dictionary-based LZ-family coding and its variants. Those methods can be interpreted as special cases under the proposed model. New compression techniques can be developed by choosing proper management policies in order to meet special encoding/decoding software or hardware requirements, or to achieve better compression performance. © 1998 Elsevier Science B.V. All rights reserved.Computer software; Encoding (symbols); Glossaries; Hierarchical systems; Information management; Mathematical models; Performance; Data coding; Data compressionjournal article10.1016/S0165-1684(98)00096-62-s2.0-0032155169