Practice topic and assessments for TREC 2010 Legal Learning Task

The following production request and seed set may be used for practice, pending release of the TREC 2010 dataset.

Please email questions to gvcormac@uwaterloo.ca.

Practice production request

Seed set and relevance assessments (column 2, 1=responsive, 0=nonresponsive)

The following dataset will be useful only for practice. For the official 2010 task, it will be necessary to download a new dataset (not yet available). The new dataset will be derived from the same source (Enron) but the format and document identifiers will not be compatible.

The trec2010 password is required to download the dataset -- it was mailed to you by NIST the "Welcome to TREC" message shortly after you filled out the TREC Application to Participate.

The TREC 2009 dataset (text format only, 850MB compressed)