摘要 |
PROBLEM TO BE SOLVED: To extract an information unit corresponding to an inputted sentence from a database with a high probability. SOLUTION: Sentence data S is formed as a simple sentence, and is stored in a text database 10. A sentence structure tree is formed by executing a morpheme analysis 4 and a sentence structure analysis 5 to the simple sentence, and sentence structure tree data 22 is formed by adding predicate attribute data of a word belonging to this tree to the individual partial tree of this sentence structure tree. Arithmetic processing 7 is executed from this sentence structure tree data 22 and the information unit in an extracting information database 9, and the information unit of a category to which the sentence data S belongs is retrieved among a plurality of information units in the extracting information database 9. COPYRIGHT: (C)2003,JPO
|