摘要 |
PROBLEM TO BE SOLVED: To extract a text with a semantic order so that a user can accurately acquire such information as meanings or contexts from large amounts of text data, without requiring the user to subjectively or newly set a subject. SOLUTION: This typical sentence analyzing device is configured to calculate category scores showing the relative relations of words to be acquired from the analysis of the internal relations of a text group in analyzing the text group, and to, on the basis of proximity relations between a designated analytic word and the other words, calculate the typicality of each text to the analytic word. COPYRIGHT: (C)2009,JPO&INPIT
|