Index of /downloads/files/eventThreading

	Name	Last modified	Size	Description
	Parent Directory		-
	eventThreading.desc	2005-02-22 15:56	11K
	eventThreading.links	2005-02-22 15:56	1.6K
	eventThreading.rel	2005-02-22 15:56	43K
	test.txt	2005-02-22 15:56	162
	train.txt	2005-02-22 15:56	156

This directory contains our experiment data in Nallapati, R., Feng, A., Peng, F., and Allan, J. , "Event Threading within News Topics" in the Proceedings of CIKM 2004 conference, pp. 446-453.

• eventThreading.rel: the story-event mapping, where the three columns in each line are TDT topic ID, story ID and our annotated event ID respectively.
• eventThreading.links: the event dependency. Each line means there is some dependency from the event indicated by the ParentID to the event numbered ChildID.
• eventThreading.desc: the description of each event.
• train.txt: the topics in our training set.
• test.txt: the topics in our test set.

The data come from TDT-2 and TDT-3. Topics numbered 200?? are TDT-2 topics, 300?? and 310?? are in TDT-3. Unfortunately, the news stories themselves are proprietary and we cannot make them available directly.

They can be obtained from the Linguistic Data Consortium at a cost that varies depending on your circumstances. Details can be found at http://www.ldc.upenn.edu/Projects/TDT2/data-release.html