Computer Laboratory

Awais Athar - Citation Sentiment Corpus

The corpus consists of 8,736 citation sentences which have been manually annotated with sentiment. These citation sentences have been extracted from the ACL Anthology Network corpus. The file format is:

Source_Paper_ID  Target_Paper_ID   Sentiment   Citation_Text

Click here to download the citation sentiment corpus corpus.

Another related resource is a list of 82 polar phrases which have been manually extracted from citation sentences in the ACL Anthology Network corpus. The file format is :

Phrase	Polarity

Polarity is 1 for positive phrases and -1 for negative ones. Spelling mistakes and OCR errors have not been corrected.

Click here to download the list of polar phrases.

For further details/citation, please see the following paper.

@InProceedings{athar:2011:SS,
  author    = {Athar, Awais},
  title     = {Sentiment Analysis of Citations using Sentence Structure-Based Features},
  booktitle = {Proceedings of the ACL 2011 Student Session},
  month     = {June},
  year      = {2011},
  address   = {Portland, OR, USA},
  publisher = {Association for Computational Linguistics},
  pages     = {81--87},
  url       = {https://www.aclweb.org/anthology/P11-3015}
}

Errata: Line 11 in Section 2 should read "and training tuning. The rest were used as the training/test set"

Source code is available at https://github.com/awaisathar/CitationSentimentClassifier