NTCIR Project

[NTCIR Home] [NTCIR Tools Home]



Mandatory tags



The tag for each document



Document identifier



Language code: CH, EN, JA, KR



Title of this news article



Issue date



Text of news article

Optional tags



Contain figures or not



Location, date or news service of the report and tags for news editors (for NewYorkTimes) *



Categorization of documents into the four distinct types; "story","multi", "advis", "other".
(for NewYorkTimes) *



Paragraph marker



Section identifier in original newspapers



Number of words in 2 bytes (for Mainichi Newspaper)

* 詳細はこちらをご参照ください:0readme.txt for the English Gigaword Third Edition http://www.ldc.upenn.edu/Catalog/docs/LDC2007T07/0readme.txt