Detecção semi-supervisionada de posicionamento em tweets baseada em regras de sentimento
Resumo
A detecção de posicionamento visa identificar automaticamente se o autor de um texto é a favor ou contra um dado alvo. O presente trabalho descreve um método semi-supervisionado de detecção de posicionamento no conteúdo textual de tweets. Um conjunto de regras é proposto para identificar posicionamento com base em opiniões positivas ou negativas a alvos direta ou indiretamente relacionados. Os tweets rotulados pelas regras são utilizados para compor um corpus de treinamento para uma abordagem supervisionada. O modelo preditivo resultante complementa a rotulação feita usando as regras. O artigo apresenta o método, e uma análise de seu desempenho quando aplicado a diferentes domínios, como candidatos políticos, mudança do clima e aborto.
Palavras-chave:
Detecção de posicionamento, Regras de sentimento, Twitter
Referências
Anand, P., Walker, M., Abbott, R., Tree, J. E. F., Bowmani, R., e Minor, M. (2011). Cats rule and dogs drool!: Classifying stance in online debate. Em Proceedings of the 2nd workshop on computational approaches to subjectivity and sentiment analysis, pgs. 1–9. Association for Computational Linguistics.
Dias, M. e Becker, K. (2016). INF-UFRGS-OPINION-MINING: Automatic generation of a training corpus for unsupervised identification of stance in tweets. Em Proceedings of the International Workshop on Semantic Evaluation - Task 6, SemEval ’16, San Diego, CA, USA. To appear.
Faulkner, A. (2014). Automated classification of stance in student essays: An approach using stance target information and the wikipedia link-based measure. Em Proceedings of the Twenty-Seventh International Florida Artificial Intelligence Research Society Conference, FLAIRS 2014, Pensacola Beach, Florida, May 21-23, 2014.
Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies, 5(1):1–167.
Mohammad, S. M., Kiritchenko, S., Sobhani, P., Zhu, X., e Cherry, C. (2016). Semeval-2016 task 6: Detecting stance in tweets. Em Proceedings of the International Workshop on Semantic Evaluation, SemEval ’16, San Diego, California.
Mohammad, S. M., Zhu, X., Kiritchenko, S., e Martin, J. (2015). Sentiment, emotion, purpose, and style in electoral tweets. Information Processing & Management, 51(4):480–499.
Rajadesingan, A. e Liu, H. (2014). Identifying users with opposing opinions in twitter debates. Em Social Computing, Behavioral-Cultural Modeling and Prediction, pgs. 153–160. Springer.
Somasundaran, S. e Wiebe, J. (2009). Recognizing stances in online debates. Em Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1-Volume 1, pgs. 226–234. Association for Computational Linguistics.
Tang, J., Nobata, C., Dong, A., Chang, Y., e Liu, H. (2015). Propagation-based sentiment analysis for microblogging data. Em Proceedings of the 2015 SIAM International Conference on Data Mining, Vancouver, BC, Canada, April 30 - May 2, 2015, pgs. 577–585.
Thomas, M., Pang, B., e Lee, L. (2006). Get out the vote: Determining support or opposition from congressional floor-debate transcripts. Em Proceedings of the 2006 conference on empirical methods in natural language processing, pgs. 327–335. Association for Computational Linguistics.
Dias, M. e Becker, K. (2016). INF-UFRGS-OPINION-MINING: Automatic generation of a training corpus for unsupervised identification of stance in tweets. Em Proceedings of the International Workshop on Semantic Evaluation - Task 6, SemEval ’16, San Diego, CA, USA. To appear.
Faulkner, A. (2014). Automated classification of stance in student essays: An approach using stance target information and the wikipedia link-based measure. Em Proceedings of the Twenty-Seventh International Florida Artificial Intelligence Research Society Conference, FLAIRS 2014, Pensacola Beach, Florida, May 21-23, 2014.
Liu, B. (2012). Sentiment analysis and opinion mining. Synthesis Lectures on Human Language Technologies, 5(1):1–167.
Mohammad, S. M., Kiritchenko, S., Sobhani, P., Zhu, X., e Cherry, C. (2016). Semeval-2016 task 6: Detecting stance in tweets. Em Proceedings of the International Workshop on Semantic Evaluation, SemEval ’16, San Diego, California.
Mohammad, S. M., Zhu, X., Kiritchenko, S., e Martin, J. (2015). Sentiment, emotion, purpose, and style in electoral tweets. Information Processing & Management, 51(4):480–499.
Rajadesingan, A. e Liu, H. (2014). Identifying users with opposing opinions in twitter debates. Em Social Computing, Behavioral-Cultural Modeling and Prediction, pgs. 153–160. Springer.
Somasundaran, S. e Wiebe, J. (2009). Recognizing stances in online debates. Em Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1-Volume 1, pgs. 226–234. Association for Computational Linguistics.
Tang, J., Nobata, C., Dong, A., Chang, Y., e Liu, H. (2015). Propagation-based sentiment analysis for microblogging data. Em Proceedings of the 2015 SIAM International Conference on Data Mining, Vancouver, BC, Canada, April 30 - May 2, 2015, pgs. 577–585.
Thomas, M., Pang, B., e Lee, L. (2006). Get out the vote: Determining support or opposition from congressional floor-debate transcripts. Em Proceedings of the 2006 conference on empirical methods in natural language processing, pgs. 327–335. Association for Computational Linguistics.
Publicado
04/10/2016
Como Citar
DIAS, Marcelo; BECKER, Karin.
Detecção semi-supervisionada de posicionamento em tweets baseada em regras de sentimento. In: SIMPÓSIO BRASILEIRO DE BANCO DE DADOS (SBBD), 31. , 2016, Salvador/BA.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2016
.
p. 40-51.
ISSN 2763-8979.
DOI: https://doi.org/10.5753/sbbd.2016.24307.
