Readme file for Summarization Data First of all, two important points you should know: 1. To obtain any of the data in this directory, you must first sign the user agreement with NII (National Institute of Informatics) You can obtain the agreement form (in English or Japanese) at http://research.nii.ac.jp/ntcir/permission/perm-ja.html http://research.nii.ac.jp/ntcir/permission/perm-en.html 2. As stated clearly in the agreement, the summarization data are summaries of the Mainichi newspaper articles, however, they are summarized for the purpose of the NTCIR Workshop 2, and The Mainichi Newspaper Co. is not responsible for their content. Note: Please also look at the details of the TSC tasks at http://oku-gw.pi.titech.ac.jp/tsc/cfp/task_description.html Econtent of files ====================================================================== Formal Run Key data F0101SENT Task A1 key data F0102FREE Task A2 key data (free summaries) F0102PART Task A2 key data (important parts) ====================================================================== Dryrun modified data 10%, 30%, 50% Important Sentence Extraction taskA1.rev ... D0101SENT 20% 40% Free Important Parts taskA2.rev ... D0102FREE D0102PART ====================================================================== Newly Added data summaries for 60 articles (1995), and 60 articles (1998) 10%, 30%, 50% Important Sentence Extraction LIST_TSCA95A1 ... TSCA95A1SENT LIST_TSCA98A1 ... TSCA98A1SENT 20% 40% Free Important Parts LIST_TSCA95A2 ... TSCA95A2FREE TSCA95A1PART LIST_TSCA98A2 ... TSCA98A2FREE TSCA98A1PART ====================================================================== ELast modified 2001.06.11 created.