摘要 |
<p><P>PROBLEM TO BE SOLVED: To highly precisely and automatically acquire highly versatile information relating to the difference of usages showing what kinds of semantic words are frequently used with synonyms or near-synonyms. <P>SOLUTION: A device A is configured of a means 1 which accepts the input of a plurality of synonyms or near-synonyms, a means 2 which extracts sentence data including input words from a corpus, a means 3 which analyzes the syntaxes of the sentence data, and extracts nouns having a grammatical relation with input words included in them, a means 4 which extracts the nodes of the nouns and the superior concepts from a thesaurus, and generates a directed graph for every input word, a means 5 which compares the plurality of directed graphs, and extracts different nodes and a means 6 which outputs the extracted nodes whose directed graphs are different as information relating to the difference of the usages of the input words. A program for making a computer function as the device A is used. <P>COPYRIGHT: (C)2007,JPO&INPIT</p> |