An Open-Domain Cause-Effect Relation Detection from Paired Nominals
We present a supervised method for detecting causal relations from text. Various kinds of dependency relations, WordNet features, Parts-of-Speech (POS) features along with several combinations of these features help to improve the performance of our system. In our experiments, we used SemEval-2010 Task #8 data sets. This system used 7954 instances for training and 2707 instances for testing from Task #8 datasets. The J48 algorithm was used to identify semantic causal relations in a pair of nominals. Evaluation result gives an overall F1 score of 85.8% of causal instances.
KeywordsCausal Relation Noun Phrase Semantic Relation Natural Language Text Receiver Operating Characteristic Space
Unable to display preview. Download preview PDF.
- 1.Blanco, E., Castell, N., Moldovan, D.: Causal Relation Extraction. In: Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC), Marrakech, Morocco (2008)Google Scholar
- 3.Hendrickx, I., Kim, S.N., Kozareva, Z., Nakov, P., Saghdha, D., Romano, L., Szpakowicz, S.: Semeval-2010 task 8: Multi-way classification of semantic relations between pairs of nominalGoogle Scholar
- 4.Sorgente, A., Vettigli, G., Mele, F.: Automatic Extraction of Cause-Effect Relations in Natural Language Text. DART@AI*IA 2013, pp. 37–48 (2013)Google Scholar
- 5.Girju, R., Moldovan, D.: Mining answers for causation questions. In: Symposium on Mining Answers from Texts and Knowledge Bases (2002)Google Scholar
- 6.Kipper-Schuler, K.: VerbNet. A broad coverage, comprehensive verb lexicon. Ph.D. thesis, University of Pennsylvania, Philadelphia, PA (2005)Google Scholar
- 7.Pal, S., Pakray, P., Das, D., Bandyopadhyay, S.: A Supervised Approach to Identify Semantic Relations from Paired Nominals. In: ACL-2010, SemEval 2010 Workshop, Uppsala, Sweden (2010)Google Scholar