Title:
文書不備検出プログラム、文書不備検出装置および文書不備検出方法
Document Type and Number:
Japanese Patent JP6606028
Kind Code:
B2
Abstract:
PROBLEM TO BE SOLVED: To automatically detect an improperness in a document described in a natural language, particularly an improperness that the contents of the document do not match the intention of a person preparing it.SOLUTION: A document improperness detection program causes a computer to realize: a function of a category determination part 11 for acquiring a document analytical model D1 that expresses the contents of a document with a digraph and category information D2 associating each document element per prescribed unit with a category, and determining a category to which each model element of the document analytical model D1 belongs; a function of a category relation determination part 12 for determining a relation among a plurality of categories in the document analytical model D1 by checking a category determination result D3 against predetermined category relation determination rules D4; and a function of an improperness detection part 13 for acquiring category relation restriction information D6 indicating a relation among the plurality of categories intended by a person preparing the document and detecting an improperness by checking a category relation determination result D5 against the category relation restriction information D6.SELECTED DRAWING: Figure 2
Inventors:
Daiki Wada
Takeshi Sumi
Takeshi Sumi
Application Number:
JP2016136816A
Publication Date:
November 13, 2019
Filing Date:
July 11, 2016
Export Citation:
Assignee:
Toshiba Corporation
International Classes:
G06F17/27; G06F8/10; G06F8/20
Domestic Patent References:
| JP2013058075A | ||||
| JP4195221A | ||||
| JP7253905A | ||||
| JP2015014827A |
Attorney, Agent or Firm:
Sakai International Patent Office