摘要 |
PURPOSE:To unnecessitate a keyword dictionary and an unwanted word dictionary by using part-of-speech information obtained by morpheme-analyzing a Japanese document, and thereby extracting a keyword. CONSTITUTION:By an input means 1, a Japanese document is inputted, an inputted document is divided into word units by a morpheme analyzing means 2, a part of speech is given to the divided word, and a keyword extracting means 3 extracts a keyword by using the part of speech. Namely, by using part-of-speech information obtained by the result analyzed by the morpheme analyzing means 2, and information of keyword identity, the keyword is extracted from in the document. And by using A composite word base being one of the keyword composition, the number of unwanted keyword candidates are decreased, and by using a proper noun constitution word being one of the keyword identity, the number of unwanted keyword candidates are decreased, and also, by using a prefix modification being one of the keyword identity, an unwanted prefix is extracted as a part of the keyword. |