| 摘要 |
PURPOSE:To obtain a high compressibility by an encoding by which a redundancy between encoded character strings is reduced, by fetching a correlation among characters in the compression-encoding of the input character strings of character information or the like which uses a universal type algorithm. CONSTITUTION:The encoded character strings are held in a dictionary 12, and the number (n) of the appearance of the coincident partial strings S1-S4 of the encoded character strings held in the dictionary 12, which start from the same leading character with a just previous character 18 to the input character string of a character inputting part 10 is retrieved. And also, the maximum length coincident partial string S4 is retrieved, an appearance number (i) indicating the order in which the just previous character of the maximum length coincident partial string appears as a starting position at the time of encoding the maximum length coincident partial string S4 is used for encoding by paring the appearance number with the coincident length, and the appearance number (i) is variable-length-encoded according the number of bits on the number (n) of appearance. The appearance number is indicated by the variable-length-code of an 'log2n' bit indicating the minimum integer more than the 'log2n'. |
