| 摘要 |
PROBLEM TO BE SOLVED: To provide an apparatus for extracting a key phrase, which enables a user to extract from documents a phrase for representing an important concept being quoted in low frequency and the concept the user requires by extracting a keyword for representing a particular category of semantic meaning in addition to the keywords conventionally extracted from documents on the basis of frequencies. SOLUTION: The apparatus for extracting the key phrases, at first, divides a document data into a word string with a data for a part of speech by parsing a morpheme of the document data, extracts the word and the word string as a candidate for the keyword on the basis of the frequencies in use of the word string as well as the word in a result of parsing the morpheme, and also extracts the word in a particular order from the result of parsing the morpheme by referring to a pattern dictionary being added with the category of semantic meaning. The apparatus determines the key phrases by using the candidate for the keyword and the result extracted by the pattern, and outputs the extracted keyword and the category of semantic meaning on the basis of the determined results. COPYRIGHT: (C)2004,JPO
|
