CONCEPT EXTRACTION SYSTEM, CONCEPT EXTRACTION METHOD, PROGRAM THEREFOR, AND STORING MEDIUM THEREOF

摘要

PROBLEM TO BE SOLVED: To provide a concept extraction technology which can extract various topics and concepts of text data accurately by repeating clustering reflecting an intention of a user.;SOLUTION: The concept extraction system extracts concepts of text data by automatically classifying a plurality of text data for a plurality of times. The system is provided with a text vector generator 1 which counts the appearance frequency of each language in the text data for every text data, and generates a matrix of each text data and each language with the appearance frequency as the matrix elements; a text vector storing part 5 which stores matrix information; a clustering part 8 which classifying every text data to a plurality of clusters based on the stored matrix information; a cluster selecting part 10 which selects a part of clusters among the plurality of clusters; a selected cluster storing part 11 which stores the selected cluster discrimination information; and a text vector amendment part 12 which amends the stored matrix information based on the cluster discrimination information.;COPYRIGHT: (C)2005,JPO&NCIPI