Automatic keyword extraction apparatus and automatic keyword extraction method

摘要

PURPOSE: To improve the quality or keywords and the retrieval precision of an information retrieval system. ;CONSTITUTION: A character string 6 is divided into character string elements (KANJI (Chinese character) character strings, KATAKANA (square form of Japanese syllabary) character strings, numeric character strings, etc.,) each of which consists of continuous characters of the same character type to generate a character string element list 7, and character strings such as numeric character strings and KATAKANA character strings which are not subjected to morpheme analysis out of these character string elements are substituted with alternative characters * or the like having corresponding grammatical information to generate a text for morpheme analysis, and this text is subjected to morpheme analysis to generate a morpheme analysis list a where at least KANJI character strings out of character string elements are divided into morphemes again. A keyword list 9 which consists or nouns or the like in character string elements and morphemes and compound words of nouns which are adjacent based on prescribed grammatical relations of nouns or the like is generated based on grammatical information in the morpheme analysis list 8.;COPYRIGHT: (C)1995,JPO