TEXT SUMMARIZATION METHOD, APPARATUS THEREOF, AND PROGRAM

摘要

PROBLEM TO BE SOLVED: To create an easily readable and easily understandable summary with a prespecified summarization rate or a length filling the limited number of characters by combining extraction of an important sentence and sentence shortening.

SOLUTION: A sentence selection unit 4 selects a sentence with the highest important order out of unselected sentences in an input document from an important sentence table 2. A sentence shortening length decision unit 5 dynamically determines a sentence shortening length for the selected sentence from a limitation length and a length of a shortened document composed of shortened sentences obtained by shortening sentences in the input document, which is stored in a shortened document table 3. A sentence shortening unit 6 shortens the selected sentence according to the sentence shortening length. A shortened sentence storage unit 7 stores the shortened sentence in a shortened document table 3 and rearranges the shortened documents on the basis of the appearance order of input documents. An output unit 8 repeats operation from sentence selection to the storage of shortened sentences till the length of the shortened document stored in the shortened document table 3 reaches a regulated rate to the limitation length, and outputs results of the repeated operation as a summary of the input document.

COPYRIGHT: (C)2011,JPO&INPIT