发明名称 |
SYSTEM AND PROGRAM FOR EXTRACTING CORRESPONDENCE RELATION BETWEEN TERMS |
摘要 |
<P>PROBLEM TO BE SOLVED: To provide technology for extracting commodity names of each enterprise based on text data, and automatically associating them to corresponding product classification. <P>SOLUTION: A system 10 for extracting correspondence relation between terms includes: a product classification dictionary 16 storing a plurality of product classifications as general names; a morphological analysis processing part 12 decomposing an input text sentence into morpheme units, and attaching a corresponding tag to a morpheme corresponding to the production classification of respective morphemes in reference to the product classification dictionary 16; an extraction rule storage part 18 storing a plurality of extraction rules each prescribing a character string pattern including the tag and a position of a character string to be extracted as the concrete commodity name belonging to the production classification attached with the tag from the character string pattern; and a relation information extraction part 14 extracting the character string present in a prescribed position inside the character string pattern matching the extraction rule inside the text sentence as the commodity name belonging to the product classification attached with the tag, and storing a combination between the product classification and the commodity name into a relation information storage part 20. <P>COPYRIGHT: (C)2011,JPO&INPIT |
申请公布号 |
JP2011103038(A) |
申请公布日期 |
2011.05.26 |
申请号 |
JP20090257213 |
申请日期 |
2009.11.10 |
申请人 |
NOMURA RESEARCH INSTITUTE LTD |
发明人 |
OSHIMA OSAMU;TAKEHARA GASUAKI;OKADA TOMOYASU |
分类号 |
G06F17/30;G06F17/27;G06Q50/22;G06Q50/24 |
主分类号 |
G06F17/30 |
代理机构 |
|
代理人 |
|
主权项 |
|
地址 |
|