Thumbs up or thumbs down?: semantic orientation applied to unsupervised classification of reviews
ACL '02 Proceedings of the 40th Annual Meeting on Association for Computational Linguistics
Recognizing contextual polarity in phrase-level sentiment analysis
HLT '05 Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing
The CoNLL-2010 shared task: learning to detect hedges and their scope in natural language text
CoNLL '10: Shared Task Proceedings of the Fourteenth Conference on Computational Natural Language Learning --- Shared Task
MAE and MAI: lightweight annotation and adjudication tools
LAW V '11 Proceedings of the 5th Linguistic Annotation Workshop
Hi-index | 0.00 |
BiasML is a novel annotation scheme with the purpose of identifying the presence as well as nuances of biased language within the subset of Wikipedia articles dedicated to service providers. Whereas Wikipedia currently uses only manual flagging to detect possible bias, our scheme provides a foundation for the automating of bias flagging by improving upon the methodology of annotation schemes in classic sentiment analysis. We also address challenges unique to the task of identifying biased writing within the specific context of Wikipedia's neutrality policy. We perform a detailed analysis of inter-annotator agreement, which shows that although the agreement scores for intra-sentential tags were relatively low, the agreement scores on the sentence and entry levels were encouraging (74.8% and 66.7%, respectively). Based on an analysis of our first implementation of our scheme, we suggest possible improvements to our guidelines, in hope that further rounds of annotation after incorporating them could provide appropriate data for use within a machine learning framework for automated detection of bias within Wikipedia.