Elijah Mayfield

Annotated Corpora

Negotiation Corpus v1.0

This corpus is comprised of 128 dialogues from the HCRC MapTask corpus. It is split into two sets: 20 conversations making up 4374 utterances which were manually coded according to the coding manual below, and 108 conversations making up 22501 utterances which were automatically coded by the system described in my 2011 ACL paper. (Coding Manual (PDF))
Last Updated August 13, 2011; contact me. Website design courtesy of Hal Daumé III.