Negochat corpus is annotated natural language human-agent dialogue corpus in the negotiation domain. The corpus was collected using Amazon Mechanical Turk following the Wizard-Of-Oz approach, where a wizard human translates the participants natural language utterances in real time into a semantic language. The corpus was introduced in the paper The Negochat Corpus of Human-agent Negotiation Dialogues.
The corpus contain 105 dialogues in the job candidate negotiation domain. The negotiation takes place between an employer and a candidate. The negotiation dialogue comprises the exchange of one or more of the following dialogue act in each utterance: Offer, Accept, Reject, Query, Greet, Quit.
If you make use of this software for research purposes, I'll appreciate citing the following:
@inproceedings{konovalov2016negochat,
title={The Negochat Corpus of Human-agent Negotiation Dialogues},
author={Konovalov, Vasily and Artstein, Ron and Melamud, Oren and Dagan, Ido},
booktitle={Proceedings of the 10th International Conference on Language Resources and Evaluation, LREC 2016},
year={2016},
pages={3141-3145},
url={www.lrec-conf.org/proceedings/lrec2016/pdf/240_Paper.pdf}
}
For any question, please contact [email protected]