| 摘要 |
PROBLEM TO BE SOLVED: To effectively divide a document by considering not only the relation of adjacent documents but also the relation of wide area concerning the document. SOLUTION: This document dividing device is provided with a language element segmenting means 1 for segmenting a language element from an electronic document with the units of a paragraph, a sentence and a line, a language inter-element relation degree evaluating means 2 for evaluating the relation degree of arbitrary two language elements with a common character or word, for example, a language inter-element relation degree matrix acquiring means 3 for acquiring the relation degrees of all the language elements by using this language inter-element relation degree evaluating means 2, and a matrix splitting means 4 for dividing the language inter-element relation degree matrix provided by this language inter-element relation degree matrix acquiring means 3 into the arrangement of partial matrixes having the high relation degree. Thus, the document is split corresponding to the splitting due to the partial matrix. |
