Title:
SYNONYM DICTIONARY CREATION DEVICE, SYNONYM DICTIONARY CREATION PROGRAM, AND SYNONYM DICTIONARY CREATION METHOD
Document Type and Number:
WIPO Patent Application WO/2019/049483
Kind Code:
A1
Abstract:
An objective of the present invention is to carry out a highly precise synonym assessment and automatically generate a synonym dictionary from a text. A morpheme analysis is carried out on a text, the text is segmented into multiple words, and a morpheme-analyzed text is obtained. A topic classification is carried out on the morpheme-analyzed text, at least one topic word associated with each topic is selected from the multiple words, and a reference word characterizing each topic is extracted from at least one of the topic words. The multiple words are multidimensionally vectorized and multiple vectors expressing each of the multiple words are obtained. At least one synonym is selected from the multiple words. At least one of the synonyms is selected such that the similarity between the vector expressing the reference word and the vector expressing each of said at least one synonym exceeds a set reference. A synonym dictionary is created wherein at least a portion of said at least one synonym has been registered.
Inventors:
YOSHIDA AKIKO (JP)
KASUBUCHI KIYOTAKA (JP)
YOSHIWA TAKAO (JP)
YAMASHITA YOSHIO (JP)
KASUBUCHI KIYOTAKA (JP)
YOSHIWA TAKAO (JP)
YAMASHITA YOSHIO (JP)
Application Number:
PCT/JP2018/024837
Publication Date:
March 14, 2019
Filing Date:
June 29, 2018
Export Citation:
Assignee:
SCREEN HOLDINGS CO LTD (JP)
International Classes:
G06F40/237
Foreign References:
| JPH11312168A | 1999-11-09 | |||
| JP2013016011A | 2013-01-24 |
Other References:
KAJI ET AL: "Extracting a topic hierarchy from a text corpus", IPSJ JOURNAL, vol. 44, no. 2, 1 February 2003 (2003-02-01), pages 405 - 420, XP055672528
HIDEAKI J ET AL.: "Automatic Synonym Acquisition Using a Context-Restricted Skip-gram Model", PROCEEDINGS OF THE TWENTY-SECOND ANNUAL MEETING OF THE ASSOCIATION FOR NATURAL LANGUAGE PROCESSING - NLP 2016; 7-11 MARCH, 2016, 2016003, pages 717 - 720, XP009519147
HIDEAKI J ET AL.: "Automatic Synonym Acquisition Using a Context-Restricted Skip-gram Model", PROCEEDINGS OF THE TWENTY-SECOND ANNUAL MEETING OF THE ASSOCIATION FOR NATURAL LANGUAGE PROCESSING - NLP 2016; 7-11 MARCH, 2016, 2016003, pages 717 - 720, XP009519147
Attorney, Agent or Firm:
YOSHITAKE Hidetoshi et al. (JP)
Download PDF: