DOI QR코드

DOI QR Code

어휘 정보와 구문 패턴에 기반한 단일 클래스 분류 모델

One-Class Classification Model Based on Lexical Information and Syntactic Patterns

  • 이현구 (강원대학교 컴퓨터정보통신공학과) ;
  • 최맹식 (강원대학교 컴퓨터정보통신공학과) ;
  • 김학수 (강원대학교 컴퓨터정보통신공학과)
  • 투고 : 2015.02.16
  • 심사 : 2015.03.31
  • 발행 : 2015.06.15

초록

관계 추출은 질의응답 및 지식확장 등에 널리 사용될 수 있는 주요 정보추출 기술이다. 정보추출에 관한 기존 연구들은 관계 범주가 수동으로 부착된 대용량의 학습 데이터를 필요로 하는 지도 학습모델을 기반으로 이루어져 왔다. 최근에는 학습 데이터 구축을 위한 인간의 노력을 줄이기 위해 원거리 감독법이 제안되었다. 그러나 원거리 감독법은 분류 문제를 해결하는데 필수적인 부정 학습 데이터를 수집하기 어렵다는 단점이 있다. 이러한 원거리 감독법의 단점을 극복하기 위해 본 논문에서는 부정 데이터 없이 학습이 가능한 단일 클래스 분류 모델을 제안한다. 입력 데이터로부터 긍정 데이터를 선별하기 위해서 제안 모델은 벡터 공간 상에서 어휘 정보와 구문 패턴에 기반한 유사도 척도를 사용하여 입력 데이터가 내부 범주에 속하는지 그렇지 않은지 판단한다. 실험에서 제안 모델은 대표적인 단일 클래스 분류 모델인 One-class SVM보다 높은 성능(0.6509 F1-점수, 0.6833 정밀도)을 보였다.

Relation extraction is an important information extraction technique that can be widely used in areas such as question-answering and knowledge population. Previous studies on relation extraction have been based on supervised machine learning models that need a large amount of training data manually annotated with relation categories. Recently, to reduce the manual annotation efforts for constructing training data, distant supervision methods have been proposed. However, these methods suffer from a drawback: it is difficult to use these methods for collecting negative training data that are necessary for resolving classification problems. To overcome this drawback, we propose a one-class classification model that can be trained without using negative data. The proposed model determines whether an input data item is included in an inner category by using a similarity measure based on lexical information and syntactic patterns in a vector space. In the experiments conducted in this study, the proposed model showed higher performance (an F1-score of 0.6509 and an accuracy of 0.6833) than a representative one-class classification model, one-class SVM(Support Vector Machine).

키워드

과제정보

연구 과제 주관 기관 : LG전자, 강원대학교

참고문헌

  1. J. Cowie, W. Lehnert, "Information extraction," communication of the ACM, Vol. 39, No. 1, pp. 80-91, 1996. https://doi.org/10.1145/234173.234209
  2. Mintz, Mike, et al., "Distant supervision for relation extraction without labeled data," Proc. of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Proc. of the AFNLP, Vol. 2, pp. 1003-1011, 2009.
  3. DBPedia ontology [Online]. Available: http://wiki.dbpedia.org/Downloads39/ (downloaded 2014, Aug. 5)
  4. The NIST ACE evaluation website [Online]. Available: http://www.itl.nist.gov/iad/mig//tests/ace/(downloaded 2015, Jan. 1)
  5. Culotta, Aron, and Jeffrey Sorensen, "Dependency tree kernels for relation extraction," Proc. of the 42nd Annual Meeting on Association for Computational Linguistics, Article No. 423, 2004.
  6. Bunescu, Razvan C., and Raymond J. Mooney, "A shortest path dependency kernel for relation extraction," Proc. of the conference on Human Language Technology and Empirical Methods in Natural Language Processing, pp. 724-731, 2005.
  7. Zhang, Min, Jie Zhang, and Jian Su, "Exploring syntactic features for relation extraction using a convolution tree kernel," Proc. of the main conference on Human Language Technology Conference of the North American Chapter of the Association of Computational Linguistics, pp. 288-295, 2006.
  8. Zhou, G., Zhang, M., Ji, D. H., and Zhu, Q., "Tree kernel-based relation extraction with context-sensitive structured parse tree information," Proc. of EMNLP-CoNLL, pp. 728-736, 2007.
  9. Tseng, Yuen-Hsien, et al., "Chinese Open Relation Extraction for Knowledge Acquisition," Proc. of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pp. 12-16, 2014.
  10. Chen, Yanping, Qinghua Zheng, and Wei Zhang, "Omni-word Feature and Soft Constraint for Chinese Relation Extraction," Proc. of the 52nd Annual Meeting of the Association for Computational Linguistics, pp. 572-581, 2014.
  11. LM Manevitz, M Yousef, "One-class SVMs for document classification," Proc. of The Journal of Machine Learning Research, Vol. 2, pp. 139-154, 2002.
  12. S Ryu, J Song, S Koo, GG Lee, "Detecting Multiple Domains from User's Utterance in Spoken Dialog System," Proc. of International Workshop Series On Spoken Dialogue Systems Technology, 2015.
  13. Apache OpenNLP [Online]. Available: https://opennlp.apache.org/ (downloaded 2014, Oct. 1)
  14. Razvan C. Bunescu and Raymond J. Mooney, "A Shortest Path Dependency Kernel for Relation Extraction," Proc. of Human Language Technology Conference and Conference on Empirical Methods in Natural Language Processing, pp. 724-731, 2005.
  15. Deepak Ravichandran and Eduard Hovy, "Learning Surface Text Patterns for a Question Answering System," Proc. of the 40th Annual Meeting on Association for Computational Linguistics, pp. 41-47, 2002.
  16. Ingo Feinerer," A text mining framework in R and its applications," pp. 156-157. 2008.
  17. CC Chang, CJ Lin, "LIBSVM: a library for support vector machines," Proc. of ACM Transactions on Intelligent Systems and Technology (TIST), Vol. 2, Issue 3, Apr. 2011, Article No. 27, 2011.