摘要 |
PURPOSE:To increase the number of words to be stored, by representing the reading of the word with a character code coincident with the statistic feature of Japanese words, and obtaining the common reading information between a word including word readings of different orders between two head characters of a character string and the above-mentioned word. CONSTITUTION:The 1st character train CS1 supplied from a character train input device 1 is converted into the 2nd character train CS2 through a character converting part 3. Then the 2nd character converting part 4 discriminates the scales of the two head characters of the train CS2 and delivers a series (the 3rd character train) CS3 formed in order smaller codes. At a word reading information retrieving part 5, the word having the word reading information coincident with the train CS3 in retrieved out of a word storage part 6, and the corresponding address AD is delivered. A forward/backward information deciding part 14 compares the forward/backward information D of the train CS2 with the forward/backward information DM of the address AD. Then the word attribute information is decided by the parts 5 and 14 and used as a word retrieving output. |