| 摘要 |
PURPOSE:To obtain the conversion rate of the round robin method with the processing efficiency of the longest coincidence method, by setting the number of conversion stages in the independent word extracting method and designating its upper limit value. CONSTITUTION:Under the command of a controlling part 1, an input/output processing part 7 sets an inputted Kana character string and its length to registers 10 and 9 respectively and transfers contents of the register 9 to a retrieval character length register 12 to start a retrieval check part 13. The check part 13 refers to an independent word dictionary 18 on a basis of contents of registers 10 and 12 and sets pertinent independent words, parts of speech, and types of conjugations and checks grammatical connection relations between independent word parts and adjunctive words in accordance with a conjugation connection table 16. If this check results in failure, 1 is subtracted from contents of the register 12 to repeat the dictionary access. When an independent word for which the check results is success is detected, a flag 11 is set, and a pair of the independent word and an adjunctive word is set to a stage-classified conversion buffer 8. The controlling part 1 starts an extraction controlling part 2 and counts up contents of a conversion stage number 5 and compares them with contents of a conversion stage number upper limit value 3 in a comparator 4; and when they coincide with each other, contents of the buffer 8 are outputted from the processing part 7. |
