| { | |
| "paper_id": "O13-1027", | |
| "header": { | |
| "generated_with": "S2ORC 1.0.0", | |
| "date_generated": "2023-01-19T08:04:03.646112Z" | |
| }, | |
| "title": "Textual Entailment Recognition Using Textual Features and SVM", | |
| "authors": [ | |
| { | |
| "first": "Tao-Hsing", | |
| "middle": [], | |
| "last": "\u5f35\u9053\u884c", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "\u8a31\u66dc\u9e92", | |
| "middle": [], | |
| "last": "Chang", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "changth@kuas.edu.tw" | |
| }, | |
| { | |
| "first": "Hsu", | |
| "middle": [], | |
| "last": "Yao-Chi", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Chung-Wei", | |
| "middle": [], | |
| "last": "\u5f35\u4e2d\u7dad", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Yao-Chuan", | |
| "middle": [], | |
| "last": "Chang", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Hsu", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "National Kaohsiung University of Applied Sciences", | |
| "location": {} | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "Hsueh-Chih", | |
| "middle": [], | |
| "last": "\u9673\u5b78\u5fd7", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Normal University", | |
| "location": { | |
| "country": "Taiwan" | |
| } | |
| }, | |
| "email": "" | |
| }, | |
| { | |
| "first": "", | |
| "middle": [], | |
| "last": "Chen", | |
| "suffix": "", | |
| "affiliation": { | |
| "laboratory": "", | |
| "institution": "Normal University", | |
| "location": { | |
| "country": "Taiwan" | |
| } | |
| }, | |
| "email": "" | |
| } | |
| ], | |
| "year": "", | |
| "venue": null, | |
| "identifiers": {}, | |
| "abstract": "The aim of this paper is to propose a system, which can automatically infer entailment relations of textual pairs. SVM is utilized as a prediction model of the system and seven features of textual pairs are employed to be input of the prediction model. The performance of this system is evaluated by dataset in CT-MC task held by RITE-2 of NTCIR. Macro-F1 of the proposed method is 46.35%.", | |
| "pdf_parse": { | |
| "paper_id": "O13-1027", | |
| "_pdf_hash": "", | |
| "abstract": [ | |
| { | |
| "text": "The aim of this paper is to propose a system, which can automatically infer entailment relations of textual pairs. SVM is utilized as a prediction model of the system and seven features of textual pairs are employed to be input of the prediction model. The performance of this system is evaluated by dataset in CT-MC task held by RITE-2 of NTCIR. Macro-F1 of the proposed method is 46.35%.", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "Abstract", | |
| "sec_num": null | |
| } | |
| ], | |
| "body_text": [ | |
| { | |
| "text": "Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing(ROCLING 2013)", | |
| "cite_spans": [], | |
| "ref_spans": [], | |
| "eq_spans": [], | |
| "section": "", | |
| "sec_num": null | |
| } | |
| ], | |
| "back_matter": [], | |
| "bib_entries": { | |
| "BIBREF0": { | |
| "ref_id": "b0", | |
| "title": "A survey of paraphrasing and textual entailment methods", | |
| "authors": [ | |
| { | |
| "first": "I", | |
| "middle": [], | |
| "last": "Androutsopoulos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [], | |
| "last": "Malakasiotis", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2010, | |
| "venue": "Journal of Artificial Intelligence Research", | |
| "volume": "38", | |
| "issue": "", | |
| "pages": "135--187", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "I. Androutsopoulos and P. Malakasiotis, \"A survey of paraphrasing and textual entailment methods,\" Journal of Artificial Intelligence Research, vol. 38, pp. 135-187, 2010.", | |
| "links": null | |
| }, | |
| "BIBREF1": { | |
| "ref_id": "b1", | |
| "title": "Recognizing textual entailment with logical inference", | |
| "authors": [ | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Bos", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Markert", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2005, | |
| "venue": "Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "628--635", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "J. Bos and K. Markert, \"Recognizing textual entailment with logical inference,\" in Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing, Association for Computational Linguistics, 2005, pp. 628-635.", | |
| "links": null | |
| }, | |
| "BIBREF2": { | |
| "ref_id": "b2", | |
| "title": "Combining specialized entailment engines for rte-4", | |
| "authors": [ | |
| { | |
| "first": "E", | |
| "middle": [], | |
| "last": "Cabrio", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Kouylekov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Magnini", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2008, | |
| "venue": "Proceedings of TAC08, 4th PASCAL Challenges Workshop on Recognizing Textual Entailment", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "E. Cabrio, M. Kouylekov and B. Magnini, \"Combining specialized entailment engines for rte-4,\" in Proceedings of TAC08, 4th PASCAL Challenges Workshop on Recognizing Textual Entailment, 2008.", | |
| "links": null | |
| }, | |
| "BIBREF3": { | |
| "ref_id": "b3", | |
| "title": "Recognizing textual entailment with tree edit distance algorithms", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Kouylekov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Magnini", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2005, | |
| "venue": "Proceedings of the First Challenge Workshop Recognizing Textual Entailment", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "17--20", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "M. Kouylekov and B. Magnini, \"Recognizing textual entailment with tree edit distance algorithms,\" in Proceedings of the First Challenge Workshop Recognizing Textual Entailment, 2005, pp. 17-20.", | |
| "links": null | |
| }, | |
| "BIBREF4": { | |
| "ref_id": "b4", | |
| "title": "Tree edit distance for recognizing textual entailment: Estimating the cost of insertion", | |
| "authors": [ | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Kouylekov", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "B", | |
| "middle": [], | |
| "last": "Magnini", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2006, | |
| "venue": "Proceedings of the PASCAL RTE-2 Challenge", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "68--73", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "M. Kouylekov and B. Magnini, \"Tree edit distance for recognizing textual entailment: Estimating the cost of insertion,\" in Proceedings of the PASCAL RTE-2 Challenge, 2006, pp. 68-73.", | |
| "links": null | |
| }, | |
| "BIBREF5": { | |
| "ref_id": "b5", | |
| "title": "The Yuntech system in NTCIR-9 RITE Task", | |
| "authors": [ | |
| { | |
| "first": "N", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Han", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Ku", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "345--348", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "N. H. Han and L. W. Ku, \"The Yuntech system in NTCIR-9 RITE Task,\" in Proceedings of the NTCIR-9 Workshop, 2011, pp. 345-348.", | |
| "links": null | |
| }, | |
| "BIBREF6": { | |
| "ref_id": "b6", | |
| "title": "NTU Textual Entailment System for NTCIR 9 RITE Task", | |
| "authors": [ | |
| { | |
| "first": "H", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Huang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [ | |
| "M" | |
| ], | |
| "last": "Haver", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Chen", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "Proceedings of the NTCIR-9 Workshop", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "349--352", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "H. H. Huang, K. C. Chang, J. M. Haver II and H. H. Chen, \"NTU Textual Entailment System for NTCIR 9 RITE Task,\" in Proceedings of the NTCIR-9 Workshop, 2011, pp. 349-352.", | |
| "links": null | |
| }, | |
| "BIBREF7": { | |
| "ref_id": "b7", | |
| "title": "A statistical parser", | |
| "authors": [ | |
| { | |
| "first": "Stanford", | |
| "middle": [], | |
| "last": "Parser", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2002, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Stanford Parser: A statistical parser, 2002, Available : http://nlp.stanford.edu/software/lex-parser.shtml", | |
| "links": null | |
| }, | |
| "BIBREF8": { | |
| "ref_id": "b8", | |
| "title": "Automated essay scoring using set of literary sememes", | |
| "authors": [ | |
| { | |
| "first": "T", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "P", | |
| "middle": [ | |
| "Y" | |
| ], | |
| "last": "Tsai", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [ | |
| "P" | |
| ], | |
| "last": "Tam", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2009, | |
| "venue": "Information: An International Interdisciplinary Journal", | |
| "volume": "12", | |
| "issue": "2", | |
| "pages": "351--357", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "T. H. Chang, C. H. Lee, P. Y. Tsai, and H. P. Tam, \"Automated essay scoring using set of literary sememes,\" Information: An International Interdisciplinary Journal, vol. 12, no. 2, pp. 351-357, 2009.", | |
| "links": null | |
| }, | |
| "BIBREF9": { | |
| "ref_id": "b9", | |
| "title": "IASL RITE System at NTCIR-10", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Shih", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [], | |
| "last": "Liu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "W", | |
| "middle": [ | |
| "L" | |
| ], | |
| "last": "Hsu", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 10th NTCIR Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "C. W. Shih, C. Liu, C. W. Lee and W. L. Hsu, \"IASL RITE System at NTCIR-10,\" in Proceedings of the 10th NTCIR Conference, Tokyo, Japan, 2013.", | |
| "links": null | |
| }, | |
| "BIBREF10": { | |
| "ref_id": "b10", | |
| "title": "Proceedings of the 9th NII Test Collection for Information Retrieval Workshop", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Akiba", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Taira", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Fujita", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Kasahara", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "M", | |
| "middle": [], | |
| "last": "Nagata", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "330--334", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Y. Akiba, H. Taira, S. Fujita, K. Kasahara and M. Nagata, \"NTTCS textual entailment recognition system for the NTCIR-9 rite,\" in Proceedings of the 9th NII Test Collection for Information Retrieval Workshop, 2011, pp. 330-334.", | |
| "links": null | |
| }, | |
| "BIBREF11": { | |
| "ref_id": "b11", | |
| "title": "The Description of the NTOU RITE System in NTCIR-10", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "J" | |
| ], | |
| "last": "Lin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Tu", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of NTCIR-10 Workshop Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "495--498", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "C. J. Lin and Y. C. Tu, \"The Description of the NTOU RITE System in NTCIR-10,\" in Proceedings of the 10th NTCIR Conference, in Proceedings of NTCIR-10 Workshop Conference,Tokyo, Japan, 2013, pp. 495-498.", | |
| "links": null | |
| }, | |
| "BIBREF12": { | |
| "ref_id": "b12", | |
| "title": "A Chinese word segmentation and POS tagging system for readability research", | |
| "authors": [ | |
| { | |
| "first": "T", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [ | |
| "T" | |
| ], | |
| "last": "Sung", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [ | |
| "T" | |
| ], | |
| "last": "Lee", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2012, | |
| "venue": "Proceedings of Paper presented at 42nd Annual Meeting of the Society for Computers in Psychology", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "T. H. Chang, Y. T. Sung and Y. T. Lee, \"A Chinese word segmentation and POS tagging system for readability research,\" in Proceedings of Paper presented at 42nd Annual Meeting of the Society for Computers in Psychology, 2012", | |
| "links": null | |
| }, | |
| "BIBREF13": { | |
| "ref_id": "b13", | |
| "title": "Automatic Chinese unknown word extraction using small-corpus-based method", | |
| "authors": [ | |
| { | |
| "first": "T", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Lee", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2003, | |
| "venue": "Proceedings of International Conference on Natural Language Processing and Knowledge Engineering", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "459--464", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "T. H. Chang and C. H. Lee, \"Automatic Chinese unknown word extraction using small-corpus-based method,\" in Proceedings of International Conference on Natural Language Processing and Knowledge Engineering, Beijing, China, 2003, pp. 459-464.", | |
| "links": null | |
| }, | |
| "BIBREF14": { | |
| "ref_id": "b14", | |
| "title": "LIBSVM: a library for support vector machines", | |
| "authors": [ | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "J" | |
| ], | |
| "last": "Lin", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2011, | |
| "venue": "ACM Transactions on Intelligent Systems and Technology (TIST)", | |
| "volume": "2", | |
| "issue": "3", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "C. C. Chang and C. J. Lin, \"LIBSVM: a library for support vector machines,\" ACM Transactions on Intelligent Systems and Technology (TIST), vol. 2, no. 3, 2011.", | |
| "links": null | |
| }, | |
| "BIBREF15": { | |
| "ref_id": "b15", | |
| "title": "KC99: A Prediction System for Chinese Textual Entailment Relation using Decision Tree", | |
| "authors": [ | |
| { | |
| "first": "T", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Hsu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [ | |
| "C" | |
| ], | |
| "last": "Hsu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [ | |
| "I" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Chang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 10th NTCIR Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "469--473", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "T. H. Chang, Y. C. Hsu, Y. C. Hsu, J. I. Chang and C. W. Chang, \"KC99: A Prediction System for Chinese Textual Entailment Relation using Decision Tree,\" in Proceedings of the 10th NTCIR Conference, Tokyo, Japan, 2013, pp. 469-473.", | |
| "links": null | |
| }, | |
| "BIBREF16": { | |
| "ref_id": "b16", | |
| "title": "Overview of the Recognizing Inference in Text (RITE-2) at the NTCIR-10", | |
| "authors": [ | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Watanabe", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "Y", | |
| "middle": [], | |
| "last": "Miyao", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "J", | |
| "middle": [], | |
| "last": "Mizuno", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Shibata", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Kanayama", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "W" | |
| ], | |
| "last": "Lee", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "J" | |
| ], | |
| "last": "Lin", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [], | |
| "last": "Shi", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "T", | |
| "middle": [], | |
| "last": "Mitamura", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "N", | |
| "middle": [], | |
| "last": "Kando", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [], | |
| "last": "Shima", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "K", | |
| "middle": [], | |
| "last": "Takeda", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of NTCIR-10 Workshop Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "Y. Watanabe, Y. Miyao, J. Mizuno, T. Shibata, H. Kanayama, C. W. Lee, C. J. Lin, , S. Shi, T. Mitamura, N. Kando, H. Shima and K. Takeda, \"Overview of the Recognizing Inference in Text (RITE-2) at the NTCIR-10,\" in Proceedings of NTCIR-10 Workshop Conference, Tokyo, Japan, 2013.", | |
| "links": null | |
| }, | |
| "BIBREF17": { | |
| "ref_id": "b17", | |
| "title": "Extracting Features for Machine Learning in NTCIR-10 RITE Task", | |
| "authors": [ | |
| { | |
| "first": "L", | |
| "middle": [], | |
| "last": "Ku", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "E", | |
| "middle": [ | |
| "T H" | |
| ], | |
| "last": "Chu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "N", | |
| "middle": [], | |
| "last": "Han", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of the 10th NTCIR Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "457--461", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "L. Ku, E. T. H. Chu and N. Han, \"Extracting Features for Machine Learning in NTCIR-10 RITE Task,\" in Proceedings of the 10th NTCIR Conference, Tokyo, Japan, 2013, pp. 457-461.", | |
| "links": null | |
| }, | |
| "BIBREF18": { | |
| "ref_id": "b18", | |
| "title": "CYUT Chinese Textual Entailment Recognition System for NTCIR-10 RITE-2", | |
| "authors": [ | |
| { | |
| "first": "S", | |
| "middle": [ | |
| "H" | |
| ], | |
| "last": "Wu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "S", | |
| "middle": [ | |
| "S" | |
| ], | |
| "last": "Yang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "L", | |
| "middle": [ | |
| "P" | |
| ], | |
| "last": "Chen", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "H", | |
| "middle": [ | |
| "S" | |
| ], | |
| "last": "Chiu", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "R", | |
| "middle": [ | |
| "D" | |
| ], | |
| "last": "Yang", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of NTCIR-10 Workshop Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "443--448", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "S. H. Wu, S. S. Yang, L. P. Chen, H. S. Chiu and R. D. Yang, \"CYUT Chinese Textual Entailment Recognition System for NTCIR-10 RITE-2,\" in Proceedings of NTCIR-10 Workshop Conference, Tokyo, Japan, 2013, pp. 443-448.", | |
| "links": null | |
| }, | |
| "BIBREF19": { | |
| "ref_id": "b19", | |
| "title": "NCCU-MIG at NTCIR-10: Using Lexical, Syntactic, and Semantic Features for the RITE Tasks", | |
| "authors": [ | |
| { | |
| "first": "W", | |
| "middle": [ | |
| "J" | |
| ], | |
| "last": "Huang", | |
| "suffix": "" | |
| }, | |
| { | |
| "first": "C", | |
| "middle": [ | |
| "L" | |
| ], | |
| "last": "Liu", | |
| "suffix": "" | |
| } | |
| ], | |
| "year": 2013, | |
| "venue": "Proceedings of NTCIR-10 Workshop Conference", | |
| "volume": "", | |
| "issue": "", | |
| "pages": "430--434", | |
| "other_ids": {}, | |
| "num": null, | |
| "urls": [], | |
| "raw_text": "W. J. Huang and C. L. Liu, \"NCCU-MIG at NTCIR-10: Using Lexical, Syntactic, and Semantic Features for the RITE Tasks,\" in Proceedings of NTCIR-10 Workshop Conference, Tokyo, Japan, 2013, pp. 430-434.", | |
| "links": null | |
| } | |
| }, | |
| "ref_entries": { | |
| "TABREF0": { | |
| "num": null, | |
| "content": "<table><tr><td>Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing (ROCLING 2013)</td></tr><tr><td>S2\uff1a\u97ed\u83dc\u539f\u7522\u65bc\u4e2d\u570b\u3002 S3\uff1a\u97ed\u83dc\u539f\u7522\u65bc\u65e5\u672c\u3002 S4\uff1a\u6c34\u83dc\u539f\u7522\u65bc\u65e5\u672c\u3002 S5\uff1a\u6c34\u83dc\u539f\u7522\u5730\u70ba\u65e5\u672c\u3002 \u8cc7\u6599\u6240\u5f97\u4e4b\u7279\u5fb5\u4f5c\u70ba\u8f38\u5165\u9805\uff0c\u4e26\u4ee5 SVM \u4f5c\u70ba\u9810\u6e2c\u6a21\u578b\u3002\u9019 7 \u9805\u7279\u5fb5\u662f\u85c9\u7531\u89c0\u5bdf\u8cc7\u6599\u6240 \u6b78\u7d0d\uff0c\u5747\u5177\u6709\u5408\u7406\u7684\u63a8\u8ad6\u89e3\u91cb\u4ee5\u53ca\u6578\u503c\u5b9a\u7fa9\u3002\u56e0\u6b64\u672c\u6587\u5c07\u5206\u6790\u6bd4\u8f03\u5728\u76f8\u540c\u7684\u9810\u6e2c\u6a21\u578b \u6642\uff0c\u4e0d\u540c\u7684\u9078\u53d6\u7279\u5fb5\u9020\u6210\u9810\u6e2c\u6587\u672c\u5c0d\u860a\u6db5\u95dc\u4fc2\u6548\u80fd\u4e0a\u7684\u5dee\u7570\u3002\u53e6\u5916\u4e5f\u6bd4\u8f03\u5728\u540c\u6a23 7 \u9805\u7279 \u5fb5\u6240\u7522\u751f\u7684\u8a13\u7df4\u8cc7\u6599\u4e0b\uff0c\u54ea\u4e00\u7a2e\u9810\u6e2c\u6a21\u578b\u5728\u9019\u500b\u554f\u984c\u4e0a\u6703\u6709\u8f03\u4f73\u7684\u8868\u73fe\u3002 \u9019\u7bc7\u8ad6\u6587\u5176\u9918\u90e8\u5206\u7d44\u7e54\u5982\u4e0b\u3002\u7b2c\u4e8c\u7bc0\u63a2\u8a0e\u76f8\u95dc\u7814\u7a76\uff0c\u5305\u62ec\u82f1\u6587\u7684\u6587\u53e5\u860a\u6db5\u7814\u7a76\u4ee5\u53ca\u8fd1\u5e74 \u4f86\u570b\u969b\u8a55\u6bd4\u4e2d\u8868\u73fe\u8f03\u4f73\u7684\u65b9\u6cd5\uff0c\u4e26\u8aaa\u660e\u8207\u6211\u5011\u63d0\u51fa\u7684\u65b9\u6cd5\u4e4b\u9593\u7684\u95dc\u4fc2\u3002\u7b2c\u4e09\u7bc0\u4ecb\u7d39\u672c\u6587 \u63d0\u51fa\u7684 7 \u9805\u7279\u5fb5\u6240\u4ee3\u8868\u7684\u610f\u7fa9\u53ca\u5b9a\u7fa9\u503c\u7684\u8a08\u7b97\u65b9\u6cd5\u3002\u53e6\u5916\u4e5f\u8aaa\u660e\u5c07\u7528\u4f86\u6bd4\u8f03\u7684\u9810\u6e2c\u6a21 \u578b\uff0c\u7b2c\u56db\u7bc0\u662f\u6bd4\u8f03\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u8207\u5176\u4ed6\u65b9\u6cd5\uff0c\u4e26\u4ee5 NTCIR RITE-2 \u63d0\u4f9b\u7684\u8a13\u7df4\u8207\u6e2c\u8a66\u8cc7 \u6599\u70ba\u4f9d\u64da\u3002\u6700\u5f8c\u8a0e\u8ad6\u672c\u7814\u7a76\u7684\u4fb7\u9650\u4ee5\u53ca\u672a\u7adf\u4e4b\u8655\uff0c\u63a2\u8a0e\u672a\u4f86\u53ef\u884c\u7684\u7814\u7a76\u65b9\u5411\u3002 \u95dc\u4fc2\u7d50\u679c\u5224\u65b7\u5169\u500b\u53e5\u5b50\u5c6c\u65bc\u54ea\u7a2e\u860a\u6db5\u95dc\u4fc2\u3002\u800c\u5404\u7a2e\u4e8c\u5143\u95dc\u4fc2\u7684\u4f9d\u64da\u5247\u5efa\u7acb\u5728\u5404\u9805\u7279\u5fb5 \u4e0a\u3002 \u4e0a\u8ff0\u7684\u7279\u5fb5\u8207\u9810\u6e2c\u6587\u672c\u860a\u6db5\u95dc\u4fc2\u90fd\u6703\u9700\u8981\u4e00\u500b\u6574\u5408\u500b\u7279\u5fb5\u7684\u5206\u985e\u6a21\u578b[11]\u3002\u652f\u6301\u5411\u91cf\u6a5f (SVM)\u662f\u6700\u666e\u904d\u7684\u5206\u985e\u6a21\u578b\uff0c\u4f8b\u5982[3]\u4ee5\u7d93\u5e38\u4f7f\u7528\u7684\u7279\u5fb5\u5305\u62ec\u6587\u5b57\u3001\u5256\u6790\u6a39\u3001\u60c5\u7dd2\u6b63\u53cd\u610f\u3001 \u540d\u8a5e\u7e2e\u5beb\u7b49\uff0c\u5c07\u6587\u53e5\u8f49\u63db\u6210\u7279\u5fb5\u5411\u91cf\uff0c\u4e26\u4f7f\u7528\u7279\u5fb5\u5411\u91cf\u63a8\u8ad6\u51fa\u860a\u6db5\u95dc\u4fc2\u3002\u53e6\u4e00\u7a2e\u5e38\u7528\u7684 \u5206\u985e\u6a21\u578b\u5247\u662f\u6c7a\u7b56\u6a39\u3002\u6c7a\u7b56\u6a39\u53ef\u7531\u5c08\u5bb6\u5efa\u69cb\u6216\u662f\u85c9\u7531\u6a5f\u5668\u5b78\u7fd2\u7684\u65b9\u5f0f\u7522\u751f\uff0cID3 \u53ef\u4ee5\u900f \u904e\u5f97\u5230\u7684\u8cc7\u8a0a\u4f86\u6700\u4f73\u5316\u6a39\u7684\u7d50\u69cb\u3002\u6839\u64da\u5148\u524d\u7684\u7814\u7a76\uff0c\u672c\u6587\u5c07\u5617\u8a66\u5229\u7528\u6587\u672c\u5c0d\u7684\u8a5e\u5f59\u3001\u8a9e \u610f\u53ca\u8a9e\u6cd5\u7279\u5fb5\u914d\u5408 SVM \u9810\u6e2c\u6a21\u578b\u63a8\u8ad6\u6587\u672c\u5c0d\u7684\u860a\u6db5\u95dc\u4fc2\u3002 \u5728 RITE-2 \u4e2d\uff0c\u652f\u6301\u5411\u91cf\u6a5f(SVM)\u4e5f\u662f\u6700\u5e38\u4f7f\u7528\u7684\u5206\u985e\u6a21\u578b[12][18-20]\u3002\u800c\u5404\u7814\u7a76\u7684\u4e3b\u8981 \u5dee\u7570\u5c31\u5728\u8f38\u5165\u7279\u5fb5\u7684\u9078\u64c7\u4e0a\u3002\u4f8b\u5982[12]\u5f37\u8abf\u4ee5\u591a\u9054 20 \u7a2e\u7684\u7279\u5fb5\u8f38\u5165 SVM \u9032\u884c\u5224\u65b7\uff1b[18] \u5247\u63d0\u5230\u4ee5\u95dc\u9375\u5b57\u7684\u5339\u914d\u53ca\u6578\u91cf\u3001\u5256\u6790\u6a39\u8a5e\u985e\u5206\u6790\u3001\u5426\u5b9a\u8a5e\u3001\u540c\u7fa9\u8a5e\u4f5c\u70ba\u7279\u5fb5\uff1b[19]\u5247\u4f7f \u7528\u6642\u9593\u8207\u6578\u5b57\u7684\u8868\u793a\u4ee5\u53ca\u5426\u5b9a\uff1b[20]\u5247\u63d0\u5230\u53e5\u6cd5\u5206\u6790\u3001\u5c08\u6709\u540d\u8a5e\u8fa8\u8a8d\u3001\u8fd1\u7fa9\u8a5e\u3001\u5e38\u7528\u8a5e \u7684\u6578\u91cf\u3001\u6587\u53e5\u9577\u5ea6\u3001\u5426\u5b9a\u8a5e\u3001\u53cd\u7fa9\u8a5e\u7684\u4f7f\u7528\u3002\u96d6\u7136\u67d0\u4e9b\u7279\u5fb5\u540c\u6642\u88ab\u4e0d\u540c\u65b9\u6cd5\u6240\u4f7f\u7528\uff0c\u4f46 \u662f\u7cfb\u7d71\u8a6e\u91cb\u8207\u8a13\u7df4\u5176\u7279\u5fb5\u7684\u65b9\u5f0f\u4ecd\u6709\u4e0d\u540c\u3002\u672c\u6587\u5c07\u767c\u5c55\u4e00\u4e9b\u7279\u5fb5\u4e26\u540c\u6a23\u4ee5 SVM \u70ba\u9810\u6e2c \u6a21\u578b\uff0c\u4ee5\u4fbf\u6bd4\u8f03\u5148\u524d\u7814\u7a76\u548c\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u7684\u7279\u5fb5\u5728\u9810\u6e2c\u860a\u6db5\u95dc\u4fc2\u4e0a\u7684\u6548\u80fd\u5dee\u7570\u3002 \u4e09\u3001\u65b9\u6cd5 \u80fd\u5728\u8b1b\u76f8\u540c\u4e8b\u7269\u7684\u6a5f\u7387\u8f03\u9ad8\u3002\u8209\u4f8b\u4f86\u8aaa\uff0c\u4ee5\u4e0b\u4e09\u500b\u53e5\u5b50 S6\u3001S7 \u548c S8 \u4e2d\u90fd\u5305\u542b\u4e86\u4e09\u500b \u540d\u8a5e\u4e14\u5728\u6558\u8ff0\u540c\u4e00\u4ef6\u4e8b\u60c5\uff0c\u56e0\u6b64\u53ef\u80fd\u662f\u96d9\u5411\u860a\u6db5\u95dc\u4fc2\uff0c\u4f8b\u5982\u6587\u672c\u5c0d(S6,S7)\u662f\u96d9\u5411\u860a\u6db5\u3002 \u4f46\u662f\u6709\u76f8\u540c\u6578\u91cf\u540d\u8a5e\u7684\u6587\u53e5\u5c0d\u4e5f\u53ef\u80fd\u56e0\u70ba\u53e5\u4e2d\u67d0\u4e9b\u6587\u5b57\u5c0e\u81f4\u53e5\u5b50\u4e92\u76f8\u77db\u76fe\uff0c\u4f8b\u5982\u6587\u672c\u5c0d (S7,S8)\u5c31\u662f\u77db\u76fe\u95dc\u4fc2\u3002 S6\uff1aH5N1 \u578b\u75c5\u6bd2\u682a\u80fd\u900f\u904e\u79bd\u985e\u50b3\u67d3\u7d66\u4eba\u9ad4 S7\uff1aH5N1 \u578b\u75c5\u6bd2\u682a\u662f\u85c9\u7531\u79bd\u985e\u50b3\u67d3\u7d66\u4eba\u9ad4 S8\uff1aH5N1 \u578b\u75c5\u6bd2\u682a\u4e26\u975e\u7531\u79bd\u985e\u50b3\u67d3\u7d66\u4eba\u9ad4 \u56e0\u6b64\u672c\u6587\u5b9a\u7fa9\u4e86\u4e00\u9805\u6587\u672c\u5c0d\u7279\u5fb5\u300c\u540d\u8a5e\u6578\u91cf\u4e00\u81f4\u6027\u300d \uff0c\u7c21\u7a31 CNN\u3002\u82e5\u6587\u672c\u5c0d\u7684\u5169\u53e5\u5b50\u540d \u8a5e\u6578\u91cf\u4e00\u81f4\uff0c\u5247\u8a72\u6587\u672c\u5c0d\u7684 CNN \u70ba 1\uff0c\u5426\u5247\u70ba-1\u3002 2. \u8a5e\u91cd\u758a\u7387\u5dee\u7570 (DRO) \u6211\u5011\u89c0\u5bdf\u5230\u7576\u4e00\u500b\u6587\u672c\u5c0d\u4e2d\u5169\u500b\u53e5\u5b50\u4f7f\u7528\u76f8\u540c\u7684\u8a5e\u8d8a\u5c11\uff0c\u8a72\u6587\u672c\u5c0d\u70ba\u7368\u7acb\u95dc\u4fc2\u7684\u6a5f\u6703\u6108 \u9ad8\u3002\u56e0\u6b64\u5c0d\u65bc\u4e00\u500b\u6587\u672c\u5c0d(S i , S j )\uff0c\u672c\u6587\u5b9a\u7fa9\u8a72\u6587\u672c\u5c0d\u7684\u300c\u9806\u5411\u8a5e\u91cd\u758a\u7387\u300d(RWF)\u53ca\u300c\u53cd \u5411\u8a5e\u91cd\u758a\u7387\u300d(RWB)\u5982\u4e0b\uff1a \u7acb\u95dc\u4fc2\u3002\u53c3\u7167 DRO \u7684\u5b9a\u7fa9\uff0c\u672c\u6587\u5b9a\u7fa9\u4e00\u500b\u6587\u672c\u5c0d\u7279\u5fb5\u300c\u8a5e\u6027\u91cd\u758a\u7387\u5dee\u7570\u300d \uff0c\u7c21\u7a31 DOP\u3002 \u5c0d\u6587\u672c\u5c0d(S i , S j )\uff0c\u8a08\u7b97 DOP \u524d\u5148\u8a08\u7b97\u8a72\u6587\u672c\u5c0d\u7684\u300c\u9806\u5411\u8a5e\u6027\u91cd\u758a\u7387\u300d(RPF)\u53ca\u300c\u53cd\u5411\u8a5e \u6027\u91cd\u758a\u7387\u300d(RPB)\u5982\u4e0b\uff1a (\u56db) \u8a5e\u610f\u7279\u5fb5 1. \u6642\u9593\u4e0d\u5c0d\u7a31(OOT) \u4f4d\u7f6e\u4e0d\u540c\u3002\u9019\u9020\u6210\u8a72\u6587\u672c\u5c0d\u662f\u77db\u76fe\u95dc\u4fc2\u3002 S19\uff1a\u7518\u8517\u662f\u88fd\u9020\u8517\u7cd6\u7684\u539f\u6599 S20\uff1a\u8517\u7cd6\u662f\u88fd\u9020\u7518\u8517\u7684\u539f\u6599 \u7136\u800c\u4e26\u975e\u6240\u6709\u5177\u6709\u5b8c\u5168\u76f8\u540c\u7684\u8a5e\u4f46\u8a5e\u5e8f\u4e0d\u540c\u7684\u6587\u672c\u5c0d\u90fd\u662f\u77db\u76fe\u95dc\u4fc2\u3002\u4ee5\u6587\u672c\u5c0d(S21, S22) \u73fe\u8c61\u537b\u672a\u5c0e\u81f4\u8a72\u6587\u672c\u5c0d\u70ba\u77db\u76fe\u95dc\u4fc2\u3002\u4e3b\u8981\u539f\u56e0\u662f\u6b64\u4e8c\u8005\u662f\u4ee5\u9023\u63a5\u8a5e\u9023\u63a5\uff0c\u4f4d\u7f6e\u4ea4\u63db\u4e26\u672a Tasks CT-BC CT-MC \u672c\u6587\u6240\u63d0\u65b9\u6cd5\u6709\u4e09\u9805\u7279\u5fb5\u8207\u77db\u76fe\u95dc\u4fc2\u6709\u95dc(WOE\u3001ENW\u3001SYN)\uff0c\u7136\u800c\u9019\u4e09\u9805\u7279\u5fb5\u8655\u7406 \u5c0e\u81f4\u8a9e\u6cd5\u7d50\u69cb\u6539\u8b8a\uff0c\u4e5f\u56e0\u6b64\u6b64\u4e8c\u53e5\u8868\u9054\u5b8c\u5168\u76f8\u540c\u7684\u610f\u601d\u3002 \u4f4e\u3002\u4e8b\u5be6\u4e0a\u77db\u76fe\u95dc\u4fc2\u7684\u63a8\u8ad6\u7121\u8ad6\u662f\u6c7a\u7b56\u6a39\u65b9\u6cd5\u8207 SVM \u90fd\u8868\u73fe\u4e0d\u4f73\u3002\u7d93\u904e\u9032\u4e00\u6b65\u5206\u6790\uff0c \u8868\u4e00\u3001\u8a55\u4f30\u8cc7\u6599\u96c6\u904b\u884c\u65bc\u4e0d\u540c\u5206\u985e\u5668\u4e4b\u7d50\u679c \u985e\u5f8c\u7684\u6574\u9ad4\u6548\u80fd\u6bd4\u5148\u524d\u63a1\u7528\u7684\u6c7a\u7b56\u6a39\u65b9\u6cd5\u66f4\u4f73\uff0c\u4f46\u662f\u63a8\u8ad6\u77db\u76fe\u95dc\u4fc2\u7684\u6b63\u78ba\u7387\u537b\u6bd4\u6c7a\u7b56\u6a39 \u70ba\u4f8b\uff0c\u5169\u53e5\u7684\u7528\u8a5e\u5b8c\u5168\u76f8\u540c\uff0c\u800c\u300c\u4f0a\u666e\u7d22\u300d\u8207 \u300c\u7f8e\u806f\u793e\u300d\u5728\u53e5\u5b50\u4e2d\u4f4d\u7f6e\u4e0d\u76f8\u540c\u3002\u9019\u500b \u8868\u4e00\u662f\u4e09\u7a2e\u4e0d\u540c\u5206\u985e\u5668\u4f7f\u7528\u76f8\u540c\u7684 7 \u9805\u7279\u5fb5\u7684\u7d50\u679c\u3002\u9019\u4e09\u7a2e\u5206\u5225\u662f\u5c08\u5bb6\u5efa\u7acb\u7684\u6c7a\u7b56\u6a39 (Decision tree)[16]\u3001\u4ee5 ID3 \u65b9\u6cd5\u81ea\u52d5\u5efa\u7acb\u7684\u6c7a\u7b56\u6a39\u4ee5\u53ca SVM\u3002\u7531\u8868\u4e00\u53ef\u4ee5\u767c\u73fe SVM \u662f \u8868\u73fe\u6700\u597d\u7684\u5206\u985e\u65b9\u5f0f\uff0cID3 \u96d6\u7136\u6bd4\u5c08\u5bb6\u5efa\u7acb Decision tree \u7684\u65b9\u6cd5\u8868\u73fe\u8f03\u4f73\uff0c\u4f46\u5176\u6574\u9ad4\u8868 \u73fe\u4ecd\u7a0d\u5fae\u843d\u5f8c SVM\u3002\u800c ID3 \u5df2\u70ba\u6700\u4f73\u5316\u4e4b\u5f8c\u7684\u7d50\u679c\uff0c\u4f46 SVM \u50c5\u4f7f\u7528 LibSVM \u7684\u9810\u8a2d \u53c3\u6578\u503c\uff0c\u82e5\u9032\u4e00\u6b65\u9032\u884c SVM \u53c3\u6578\u6700\u4f73\u5316\uff0c\u5169\u8005\u6703\u6709\u66f4\u660e\u986f\u7684\u5dee\u8ddd\u3002 \u4e94\u3001\u8a0e\u8ad6\u8207\u672a\u4f86\u5de5\u4f5c \u5f9e\u5be6\u9a57\u7d50\u679c\u53ef\u4ee5\u770b\u51fa\u672c\u6587\u6240\u63d0 7 \u9805\u7279\u5fb5\u53ef\u4ee5\u7528\u4ee5\u5340\u5225\u6587\u672c\u5c0d\u860a\u6db5\u95dc\u4fc2\u3002\u800c\u4f7f\u7528 SVM \u5206 \u5176\u4e2d P k 0.2\u3002 \u6211\u5011\u5b9a\u7fa9\u4e00\u500b\u6587\u672c\u5c0d\u7279\u5fb5\u300c\u4f7f\u7528\u540c\u7fa9\u8a5e\u300d \uff0c\u7c21\u7a31 SYN\uff0c\u5176\u503c\u5b9a\u7fa9\u5982\u4e0b\u3002\u82e5\u4e00\u500b\u6587\u672c\u5c0d\u4e2d \u76f8\u5c0d\u61c9\u7684\u4f4d\u7f6e\u4f7f\u7528\u5c11\u6578\u4e0d\u540c\u8a5e\u5f59\uff0c\u800c\u5728\u540c\u4e00\u4f4d\u7f6e\u7684\u8a5e\u5f59\u4e92\u70ba\u540c\u7fa9\u8a5e\uff0c\u4e14\u5169\u53e5\u7684\u8a5e\u6027\u9806\u5e8f \u4e5f\u76f8\u540c\uff0c\u8a72\u6587\u672c\u5c0d\u7684 SYN \u503c\u70ba 1\uff0c\u53cd\u4e4b\u70ba-1\u3002 S17\uff1a\u82e5\u671b\u4fdd\u797f\u4e8c\u4e16\u662f\u6559\u5ef7\u9818\u5c0e\u4eba S18\uff1a\u82e5\u671b\u4fdd\u797f\u4e8c\u4e16\u662f\u68b5\u8ae6\u5ca1\u9818\u5c0e\u4eba 4. \u8a5e\u5e8f\u4ea4\u63db(WOE) \u548c SYN \u76f8\u53cd\uff0c\u6709\u4e9b\u6587\u672c\u5c0d\u4e2d\u4f7f\u7528\u4e86\u5b8c\u5168\u76f8\u540c\u7684\u8a5e\uff0c\u4f46\u53ea\u662f\u56e0\u9806\u5e8f\u4e0d\u540c\uff0c\u5c0e\u81f4\u5169\u53e5\u7684\u610f \u601d\u5b8c\u5168\u76f8\u53cd\u3002\u548c ENW \u4e0d\u540c\u7684\u662f\uff0c\u9019\u6a23\u7684\u6587\u672c\u5c0d\u4e2d\u4e26\u6c92\u6709\u5426\u5b9a\u8a5e\u3002\u4f8b\u5982\u4e0b\u5217\u6587\u672c\u5c0d (S19,S20)\uff0c\u96d6\u7136\u4f7f\u7528\u7684\u8a5e\u5b8c\u5168\u4e00\u6a23\uff0c\u4f46\u610f\u601d\u5b8c\u5168\u76f8\u53cd\uff0c\u539f\u56e0\u662f\u7518\u8517\u548c\u8517\u7cd6\u5728\u53e5\u6cd5\u529f\u80fd\u7684 \u65b9\u6cd5\u6b78\u985e\u70ba\u96d9\u5411(Bidirection)\u3001\u6b63\u5411(Forward)\u3001\u77db\u76fe(Contraction)\u548c\u7368\u7acb(Independent)\u56db \u7a2e\u95dc\u4fc2\u4e4b\u4e00\u3002\u6b64\u5be6\u9a57\u5c07\u6bd4\u8f03\u4e0d\u540c\u9810\u6e2c\u6a21\u578b\u5728\u9019\u4e9b\u4efb\u52d9\u4e2d\u7684\u8868\u73fe\u3002\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u5c07\u7a31\u70ba KC99-SVM\u3002 \u5728\u5be6\u4f5c\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u6642\uff0c\u8003\u616e\u5728\u5206\u6790\u6587\u672c\u5c0d\u7279\u5fb5\u6642\u672a\u77e5\u8a5e\u7684\u7279\u6b8a\u9700\u8981\uff0c\u672c\u6587\u4ee5[13]\u63d0\u51fa \u7684 WeCAn \u7cfb\u7d71\u70ba\u57fa\u790e\uff0c\u52a0\u4ee5\u4fee\u6539\u5f8c\u5c0d\u6587\u672c\u5c0d\u53e5\u5b50\u9032\u884c\u65b7\u8a5e\u8207\u8a5e\u6027\u6a19\u8a18\u3002\u8a72\u7cfb\u7d71\u88ab\u4fee\u6539 \u70ba\u5148\u81f3 wiki \u8490\u96c6\u5c08\u6709\u540d\u8a5e\uff0c\u518d\u63a1\u7528 SPLR \u65b9\u6cd5[14]\u63d0\u9ad8\u7cfb\u7d71\u8fa8\u8b58\u672a\u77e5\u8a5e\u7684\u80fd\u529b\u3002\u53e6\u5916\uff0c \u672c\u6587\u4e5f\u5229\u7528\u898f\u5247\u5f0f\u7684\u65b9\u6cd5\u4f86\u5c07\u6578\u503c\u8cc7\u6599\u8f49\u63db\u6210\u76f8\u540c\u7684\u683c\u5f0f\u3002\u800c\u5c0d\u65bc\u540c\u7fa9\u8a5e\u7684\u5224\u65b7\uff0c\u672c\u6587 \u5c07\u53ef\u80fd\u662f\u540c\u7fa9\u8a5e\u7684\u8a5e\u9001\u81f3 Google \u82f1\u8b6f\uff0c\u82e5\u986f\u793a\u76f8\u540c\u7684\u82f1\u6587\u8a5e\u5f59\u5247\u662f\u70ba\u540c\u7fa9\u8a5e\u3002\u53e6\u5916\u5426 \u5b9a\u8a5e\u5247\u662f\u4ee5\u5217\u8868\u8f14\u4ee5\u898f\u5247\u5f0f\u65b9\u5f0f\u8655\u7406\u3002\u53e6\u5916\u5728\u4f7f\u7528 LibSVM \u6642\uff0c\u6211\u5011\u5747\u4f7f\u7528\u8a72\u7cfb\u7d71\u9810\u8a2d \u503c\u5efa\u69cb\u672c\u6587\u6240\u4f7f\u7528\u7684 SVM \u6a21\u578b\uff0c\u4e26\u672a\u9032\u4e00\u6b65\u9032\u884c\u53c3\u6578\u6700\u4f73\u5316\u3002 \u8005\u8868\u73fe\u5dee\u8ddd\u975e\u5e38\u5c0f\u3002\u7531\u65bc\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u63a1\u7528\u7684\u67b6\u69cb\u8f03\u55ae\u7d14\uff0c\u63a1\u7528\u7684\u6307\u6a19\u4e5f\u8f03\u5c11\uff0c\u56e0\u6b64\u53ef \u80fd\u5728\u67d0\u4e9b\u61c9\u7528\u4e0a\u6703\u66f4\u9069\u5408\u4f5c\u70ba\u89e3\u6c7a\u65b9\u6848\u3002 \u8868\u4e09\u3001\u8207 RITE2 \u4efb\u52d9\u4e2d Macro-F1 \u6700\u9ad8\u7684 IASL \u4e4b\u6bd4\u8f03 Tasks Indicator CT-BC CT-MC Y N B F C I Macro-F1 F1 71.66 62.63 52.35 64.63 29.90 38.41 IASL[17] \u5c0d\u6587\u672c\u5c0d(S1, \u9019\u7bc7\u8ad6\u6587\u7684\u76ee\u7684\u662f\u63d0\u51fa\u4e00\u500b\u80fd\u5224\u65b7\u6587\u672c\u5c0d\u860a\u6db5\u95dc\u4fc2\u7684\u7cfb\u7d71\u3002\u672c\u7cfb\u7d71\u4e3b\u8981\u4f7f\u7528 7 \u9805\u7531\u89c0\u5bdf \u679c\u3002[7]\u5247\u63d0\u51fa\u4e00\u7a2e\u57fa\u65bc\u8a9e\u6cd5\u5206\u6790\u7684\u65b9\u6cd5\u3002\u9996\u5148\uff0c\u4ed6\u5011\u4f7f\u7528 stanford parser[8]\u5206\u6790\u6587\u53e5\u7684 \u8a9e\u6cd5\u6a39\uff0c\u4e26\u6a19\u793a\u51fa\u4e3b\u8981\u7684\u52d5\u8a5e\u8207\u540d\u8a5e\u3002\u63a5\u8457\u5728\u5206\u6790\u4e0d\u540c\u985e\u578b\u7684\u4e3b\u8981\u52d5\u8a5e\u8207\u540d\u8a5e\u5f8c\u6b78\u7d0d\u51fa \u5e7e\u7a2e\u4e3b\u8981\u7279\u5fb5\uff0c\u6700\u5f8c\u4f7f\u7528\u9019\u4e9b\u7279\u5fb5\u4f86\u8a08\u7b97\u6587\u53e5\u4e4b\u9593\u7684\u53e5\u6cd5\u76f8\u4f3c\u5ea6\u3002\u5be6\u9a57\u7d50\u679c\u8b49\u660e[7]\u7684 \u6548\u80fd\u8f03\u53ea\u4f7f\u7528\u6dfa\u5c64\u7279\u5fb5\u63a8\u8ad6\u860a\u6db5\u95dc\u4fc2\u6709\u66f4\u597d\u7684\u8868\u73fe\u30022013 \u5e74\u6240\u8209\u8fa6\u7684 RITE2 \u4e2d\uff0c\u6548\u80fd \u5426\u885d\u7a81\u3001\u7b2c\u4e00\u53e5\u662f\u5426\u63a8\u8ad6\u7b2c\u4e8c\u53e5\u3001\u7b2c\u4e8c\u53e5\u662f\u5426\u63a8\u8ad6\u7b2c\u4e00\u53e5\u7b49\u3002\u7531\u5169\u500b\u53e5\u5b50\u5448\u73fe\u7684\u9019\u4e09\u7a2e \u7684\u6a5f\u6703\u6108\u9ad8\u3002\u9019\u662f\u56e0\u70ba\u540d\u8a5e\u662f\u7528\u4f86\u8868\u793a\u67d0\u4e9b\u4e8b\u7269\uff0c\u800c\u5169\u53e5\u5b50\u540d\u8a5e\u6578\u91cf\u76f8\u540c\u4ee3\u8868\u5169\u53e5\u5b50\u53ef \u5f9e DRO \u9032\u4e00\u6b65\u5ef6\u4f38\uff0c\u6211\u5011\u5047\u8a2d\u7576\u5169\u500b\u53e5\u5b50\u4f7f\u7528\u7684\u76f8\u540c\u8a5e\u6027\u8d8a\u5c11\uff0c\u6587\u672c\u5c0d\u5c31\u8d8a\u53ef\u80fd\u70ba\u7368 (Bidirection)\u6216\u77db\u76fe(Contraction)\u5169\u7a2e\u95dc\u4fc2\u4e4b\u4e00\u3002\u5728 CT-MC \u4efb\u52d9\u4e2d\uff0c\u6587\u672c\u5c0d\u61c9\u8a72\u88ab\u9810\u6e2c \u8868\u4e09\u662f\u6bd4\u8f03\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u8207 RITE2 \u4e2d\u6548\u679c\u6700\u597d\u7684 IASL[17]\u65b9\u6cd5\u6bd4\u8f03\u3002\u7531\u8868\u4e09\u53ef\u77e5\uff0c\u5169 \u9806\u5e8f\u8207\u5176\u8a5e\u6027\u5b8c\u5168\u76f8\u540c\u3002\u800c\u300c\u6559\u5ef7\u300d\u8207\u300c\u68b5\u8ae6\u5ca1\u300d\u662f\u540c\u7fa9\u8a5e\uff0c\u56e0\u6b64\u8a72\u6587\u672c\u5c0d\u70ba\u96d9\u5411\u95dc\u4fc2\u3002 \u6700\u4f73\u7684 IASL[10]\u63d0\u51fa\u4ee5\u4e8c\u5143\u95dc\u4fc2\u5206\u985e\u7684\u6982\u5ff5\u3002[10]\u8a8d\u70ba\u5169\u500b\u53e5\u5b50\u95dc\u4fc2\u6709\u4e09\u7a2e\uff1a\u53e5\u5b50\u9593\u662f (\u4e8c) \u8a5e\u5f59\u7279\u5fb5 \u6211\u5011\u89c0\u5bdf\u5230\u4e00\u500b\u73fe\u8c61: \u7576\u5169\u500b\u53e5\u5b50\u4e2d\u7684\u540d\u8a5e\u6578\u91cf\u4e00\u6a23\u6642\uff0c\u9019\u7d44\u6587\u672c\u5c0d\u70ba\u96d9\u5411\u53ca\u77db\u76fe\u95dc\u4fc2 (S17,S18)\u70ba\u4f8b\uff0c\u5169\u53e5\u53ea\u6709\u5728\u300c\u6559\u5ef7\u300d\u8207\u300c\u68b5\u8ae6\u5ca1\u300d\u7684\u4f4d\u7f6e\u4f7f\u7528\u4e0d\u540c\u8a5e\u5f59\uff0c\u4f46\u8a5e\u5f59\u7684\u51fa\u73fe CT-BC \u8207 CT-MC \u5169\u9805\u4efb\u52d9\uff0c\u5728 CT-BC \u4efb\u52d9\u4e2d\uff0c\u6587\u672c\u5c0d\u53ea\u9700\u88ab\u9810\u6e2c\u65b9\u6cd5\u6b78\u985e\u70ba\u96d9\u5411 1. \u8a5e\u6027\u91cd\u758a\u7387\u5dee\u7570(DOP) \u53e5\u53ef\u80fd\u662f\u63cf\u8ff0\u540c\u6a23\u4e8b\u4ef6\u6216\u4e8b\u5be6\u7684\u5169\u500b\u4e0d\u540c\u8aaa\u6cd5\uff0c\u56e0\u6b64\u53ef\u80fd\u662f\u96d9\u5411\u860a\u6db5\u95dc\u4fc2\u3002\u4ee5\u6587\u672c\u5c0d \u7684 development \u5b50\u8cc7\u6599\u96c6\uff0c\u6e2c\u8a66\u8cc7\u6599\u70ba\u4efb\u52d9\u8cc7\u6599\u96c6\u4e2d\u7684 formal run \u5b50\u8cc7\u6599\u96c6\u3002RITE-2 \u6709 Recall 86.42 39.55 73.51 75.00 6.14 42.71 1. \u540d\u8a5e\u6578\u91cf\u4e00\u81f4\u6027(CNN) (\u4e09) \u8a5e\u6cd5\u7279\u5fb5 \u6709\u4e9b\u6587\u672c\u5c0d\u7684\u5169\u500b\u53e5\u5b50\u4f7f\u7528\u7684\u8a5e\u5f59\u5927\u90e8\u5206\u76f8\u540c\uff0c\u4e14\u8a5e\u5f59\u51fa\u73fe\u9806\u5e8f\u8207\u8a5e\u5f59\u8a5e\u6027\u4e5f\u90fd\u5b8c\u5168\u76f8 \u540c\uff0c\u53ea\u6709\u5c11\u90e8\u5206\u76f8\u5c0d\u61c9\u4f4d\u7f6e\u7684\u8a5e\u5f59\u4e0d\u540c\u3002\u82e5\u9019\u4e9b\u4e0d\u540c\u7684\u8a5e\u5f59\u662f\u540c\u7fa9\u8a5e\uff0c\u5247\u8a72\u6587\u672c\u5c0d\u7684\u5169 KC99-SVM Precision 62.96 70.67 53.11 55.03 50.00 58.29 46.35 \u7570\u3002\u5be6\u9a57\u8cc7\u6599\u662f\u7531 NTCIR-10 \u4e2d RITE-2 \u7684\u4efb\u52d9\u8cc7\u6599\u96c6\u7372\u5f97\u3002\u8a13\u7df4\u8cc7\u6599\u70ba\u4efb\u52d9\u8cc7\u6599\u96c6\u4e2d F1 72.78 50.72 61.67 63.48 10.94 49.30 \u672c\u5be6\u9a57\u4e3b\u8981\u76ee\u7684\u5728\u89c0\u5bdf\u672c\u6587\u6240\u63d0\u7279\u5fb5\u662f\u5426\u5177\u6709\u826f\u597d\u9810\u6e2c\u6027\u3001\u4ee5\u53ca\u4e0d\u540c\u9810\u6e2c\u6a21\u578b\u9020\u6210\u7684\u5dee Recall 14.61 56.97 77.48 77.13 0.88 41.67 S1\uff1a\u97ed\u83dc\u539f\u7522\u65bc\u4e2d\u570b\uff0c\u662f\u5e38\u898b\u7684\u852c\u83dc\u4e4b\u4e00\u3002 \u9664\u4e86\u4ee5\u8a5e\u5f59\u548c\u8a9e\u610f\u7279\u5fb5\u7684\u76f8\u4f3c\u7a0b\u5ea6\u63a8\u8ad6\u860a\u6db5\u95dc\u4fc2\u4e4b\u5916\uff0c\u9084\u6709\u4f7f\u7528\u5256\u6790\u6a39(parsing tree)\u4f86 \u5206\u6790\u53e5\u6cd5\u7d50\u69cb\u4ee5\u63a8\u8ad6\u860a\u6db5\u95dc\u4fc2\u7684\u65b9\u6cd5[3-5]\u3002\u9019\u4e9b\u65b9\u6cd5\u90fd\u662f\u5148\u4f7f\u7528 parser \u628a\u6587\u53e5\u4ee5\u6a39\u72c0\u7d50 \u69cb\u8868\u793a\uff0c\u800c [3]\u904b\u7528 linear distance \u8207 tree edit distance \u7b49\u65b9\u6cd5\u8a08\u7b97\u6587\u53e5\u5dee\u7570\u3002[4]\u548c[5]\u5247 \u5c07\u5169\u500b\u53e5\u5b50\u7684\u6a39\u72c0\u7d50\u69cb\u5728\u7d93\u904e\u6578\u6b21\u7684\u63d2\u5165\u3001\u522a\u9664\u3001\u4ee3\u63db\u5f8c\u5c07\u6a39\u72c0\u5716\u8abf\u6574\u6210\u76f8\u540c\u7684\u5716\uff0c\u800c \u5176\u904e\u7a0b\u4e2d\u63d2\u5165\u3001\u522a\u9664\u3001\u4ee3\u63db\u7684\u6b21\u6578\u7a31\u70ba\u6a39\u8ddd(tree distance)\uff0c\u53ef\u7528\u4f86\u7576\u4f5c\u6a39\u72c0\u5716\u4e4b\u9593\u7684\u5dee \u7570\u6a19\u6e96\u3002\u9019\u4e9b\u7814\u7a76\u5229\u7528\u9019\u500b\u5dee\u7570\u6027\u4f86\u5224\u65b7\u6587\u53e5\u7684\u860a\u6db5\u95dc\u4fc2\u3002 \u7a76\u4e2d\uff0c\u672a\u77e5\u8a5e\u662f\u4e00\u500b\u9700\u8981\u8655\u7406\u7684\u554f\u984c\uff0c\u56e0\u70ba\u8a31\u591a\u5c08\u6709\u540d\u8a5e\u5927\u91cf\u51fa\u73fe\u5728\u4ee5\u77e5\u8b58\u70ba\u4e3b\u7684\u6587\u672c \u5c0d\u4e2d\u3002\u4f46\u662f\u7531\u65bc\u4e2d\u6587\u53e5\u4e2d\u8a5e\u5f59\u9593\u6c92\u6709\u7a7a\u767d\u5206\u9694\uff0c\u6240\u4ee5\u8981\u8fa8\u8b58\u672a\u77e5\u8a5e\u662f\u5f88\u56f0\u96e3\u7684\u5de5\u4f5c\u3002 \u53e6\u5916\u8cc7\u6599\u683c\u5f0f\u4e0d\u4e00\u81f4\u7684\u554f\u984c\uff0c\u4e5f\u5e38\u767c\u751f\u5728\u6587\u672c\u5c0d\u4e2d\u3002\u4ee5\u4e0b\u4e09\u7a2e\u662f\u5e38\u898b\u7684\u72c0\u6cc1\uff1a 1. \u7528\u4e0d\u540c\u7684\u65b9\u5f0f\u8868\u793a\u76f8\u540c\u7684\u8cc7\u6599\uff0c\u4f8b\u5982\u4e00\u534a\u30011/2\u30010.5 2. \u7e2e\u5beb\uff0c\u4f8b\u5982 2003 \u5e74\u300103 \u5e74 3. \u55ae\u4f4d\u8f49\u63db\uff0c\u4f8b\u5982 1kg\u30011000g 2. \u5b58\u5728\u5426\u5b9a\u8a5e(ENW) \u5728\u4e00\u4e9b\u6587\u672c\u5c0d\u4e2d\uff0c\u5169\u53e5\u8a71\u6709\u8457\u9ad8\u76f8\u4f3c\u5ea6\u4f46\u662f\u5169\u53e5\u8a71\u8868\u793a\u7684\u610f\u601d\u537b\u56e0\u70ba\u5426\u5b9a\u8a5e\u7684\u51fa\u73fe\u800c\u9020 \u6210\u77db\u76fe\u3002\u4ee5\u6587\u672c\u5c0d(S15,S16)\u70ba\u4f8b\uff0c\u53e5\u5b50 S15 \u6bd4 S16 \u591a\u51fa\u4e86\u5426\u5b9a\u8a5e\u300c\u4e0d\u6703\u300d \uff0c\u4f7f\u5f97\u8a72\u6587\u672c \u5c0d\u70ba\u77db\u76fe\u95dc\u4fc2\u3002\u56e0\u6b64\u672c\u6587\u4f7f\u7528\u4e86\u7279\u5fb5\u300c\u5b58\u5728\u5426\u5b9a\u8a5e\u300d \uff0c\u7c21\u7a31 ENW\uff0c\u5176\u503c\u5b9a\u7fa9\u70ba\uff1a\u6587\u672c\u5c0d \u4e2d\u82e5\u6709\u4e00\u53e5\u51fa\u73fe\u5426\u5b9a\u8a5e\u800c\u53e6\u4e00\u53e5\u5247\u7121\uff0c\u5247\u8a72\u6587\u672c\u5c0d\u7684 ENW \u70ba 1\uff0c\u5426\u5247\u70ba-1\u3002 S15\uff1a\u963f\u65af\u5339\u9748\u4e0d\u6703\u5f15\u8d77\u4e0d\u826f\u53cd\u61c9 \u5927\u5e45\u63d0\u9ad8\u6b63\u78ba\u7387\u7684\u9014\u5f91\u4e4b\u4e00\u3002 \u8a8c\u8b1d \u672c\u6587\u4f5c\u8005\u611f\u8b1d\u570b\u79d1\u6703\u8a08\u756b\u7de8\u865f NSC 102-2511-S-151-002 \u7684\u652f\u6301\uff0c\u540c\u6642\u4e5f\u611f\u8b1d\u6559\u80b2\u90e8\u53ca\u570b \u7acb\u53f0\u7063\u5e2b\u7bc4\u5927\u5b78\u300c\u9081\u5411\u9802\u5c16\u5927\u5b78\u8a08\u756b\u300d\u7684\u652f\u6301\u3002 \u8868\u4e8c\u662f\u6bd4\u8f03\u63a1\u7528\u672c\u6587\u6240\u63d0 7 Tasks CT-BC CT-MC 2011 \u5e74\u7531 NTCIR-9 \u6240\u8209\u8fa6 RITE \u6587\u53e5\u860a\u6db5\u63a8\u8ad6\u7684\u4efb\u52d9\u4e2d\uff0c [6]\u63a1\u7528\u5305\u62ec\u53e5\u5b50\u9577\u5ea6\u3001\u5167\u6587 \u95dc\u9375\u5b57\u91cd\u8907\u7387\u3001\u95dc\u9375\u5b57\u91cd\u8907\u7684\u6578\u91cf\u8207\u8a5e\u6027\u7b49\u7b49\u6dfa\u5c64\u7279\u5fb5\uff0c\u85c9\u6b64\u5206\u8fa8\u51fa\u5169\u500b\u53e5\u5b50\u4e4b\u9593\u7684\u5dee \u7570\u4f86\u5224\u65b7\u6587\u672c\u860a\u6db5\u95dc\u4fc2\u3002\u5be6\u9a57\u7d50\u679c\u8b49\u660e\uff0c\u53ea\u4f7f\u7528\u6dfa\u5c64\u7279\u5fb5\u63a8\u8ad6\u860a\u6db5\u95dc\u4fc2\u4e5f\u6709\u826f\u597d\u7684\u6548 \u96d6\u7136\u4ee5\u4e0a\u9019\u4e9b\u554f\u984c\u983b\u7e41\u51fa\u73fe\u5728\u5404\u7a2e\u6587\u672c\u5c0d\u4e2d\uff0c\u4f46\u662f\u683c\u5f0f\u537b\u591a\u662f\u5e38\u898b\u7684\u5e7e\u7a2e\u3002\u5728\u5be6\u9a57\u90e8\u5206 \u672c\u6587\u6703\u9032\u4e00\u6b65\u8aaa\u660e\u5728\u524d\u8655\u7406\u968e\u6bb5\u6211\u5011\u6240\u63a1\u7528\u7684\u5de5\u5177\u4ee5\u53ca\u524d\u8ff0\u554f\u984c\u7684\u89e3\u6c7a\u7b56\u7565\u3002 S16\uff1a\u963f\u65af\u5339\u9748\u53ef\u80fd\u5f15\u8d77\u4e0d\u826f\u53cd\u61c9 3. \u4f7f\u7528\u540c\u7fa9\u8a5e(SYN) NTOUA-03[17] Precision 28.81 35.89 50.43 55.00 5.26 70.59 44.80 \u56db\u3001\u5be6\u9a57 Indicator Y N B F C I Macro-F1 F1 19.39 44.04 61.10 64.21 1.50 52.40 \u53c3\u8003\u6587\u737b</td></tr><tr><td>\u4e8c\u3001\u6587\u737b\u56de\u9867 \u95dc\u65bc\u82f1\u8a9e\u6587\u672c\u7684\u860a\u6db5\u63a8\u8ad6\u5df2\u7d93\u6709\u5f88\u591a\u76f8\u95dc\u7814\u7a76\u3002[1]\u904b\u7528\u5b57\u9762\u4e0a\u7684\u76f8\u4f3c\u5ea6\u4f86\u5224\u65b7\u860a\u6db5\u95dc \u4fc2\u3002\u5982\u679c\u662f\u76f8\u540c\u7684\u8a5e\u5f59\uff0c\u6b64\u65b9\u6cd5\u78ba\u5be6\u53ef\u4ee5\u7cbe\u6e96\u7684\u5224\u65b7\u51fa\u860a\u6db5\u95dc\u4fc2\uff0c\u4f46\u662f\u9019\u6a23\u7684\u65b9\u6cd5\u5728\u8655 \u7406\u540c\u7fa9\u8a5e\u7684\u6587\u672c\u6642\u5bb9\u6613\u932f\u5224\u3002[2]\u63d0\u51fa\u7684\u300c\u6dfa\u5c64\u8a9e\u610f\u7279\u5fb5\u300d(Shallow Semantic Features)\u7684 \u7be9\u9078\u6cd5\u5247\u53ef\u4ee5\u89e3\u6c7a\u9019\u500b\u554f\u984c\u3002[2]\u4f7f\u7528 WordNet \u4f5c\u70ba\u80cc\u666f\u77e5\u8b58\uff0c\u89e3\u91cb\u4e0d\u540c\u7684\u8a5e\u5f59\u662f\u76f8\u540c \u6216\u76f8\u53cd\u7684\u8a5e\u610f\u3002\u4f8b\u5982\u300c\u5147\u624b\u300d \u3001 \u300c\u53d7\u5bb3\u8005\u300d\u8207\u300c\u8b00\u6bba\u300d\u76f8\u95dc\uff0c\u4f46\u300c\u5147\u624b\u300d\u662f\u300c\u8b00\u6bba\u300d\u7684 \u884d\u751f\u5b57\uff0c\u800c\u300c\u53d7\u5bb3\u8005\u300d\u8207\u300c\u5147\u624b\u300d\u662f\u53cd\u7fa9\u5b57\u3002 \u672c\u6587\u4f7f\u7528\u4e86\u4e03\u500b\u6587\u672c\u5c0d\u7279\u5fb5\u9810\u6e2c\u6587\u672c\u5c0d\u7684\u860a\u6db5\u95dc\u4fc2\u3002\u4e03\u9805\u7279\u5fb5\u5927\u81f4\u5206\u985e\u6210\u8a5e\u5f59\u3001\u8a9e\u610f\u53ca \u8a9e\u6cd5\u4e09\u7a2e\u3002\u6b64\u5916\uff0c\u4e2d\u6587\u6587\u672c\u5c0d\u5728\u8a08\u7b97\u7279\u5fb5\u524d\u9700\u8981\u65b7\u8a5e\u53ca\u8a5e\u6027\u6a19\u8a18\u7b49\u524d\u8655\u7406\u5de5\u4f5c\uff0c\u4ee5\u4fbf\u5f8c \u7e8c\u6f14\u7b97\u6cd5\u4f7f\u7528\u3002\u524d\u8655\u7406\u4ee5\u53ca\u4e03\u500b\u7279\u5fb5\u7684\u7d30\u7bc0\u5728\u4e0b\u5217\u5404\u5c0f\u7bc0\u8aaa\u660e\u3002 (\u4e00) \u524d\u8655\u7406 \u7531\u65bc\u4e2d\u6587\u9593\u8a5e\u8207\u8a5e\u4e4b\u9593\u6c92\u6709\u7a7a\u767d\u5206\u9694\uff0c\u56e0\u6b64\u4e2d\u6587\u6587\u672c\u5c0d\u5fc5\u9808\u5148\u9032\u884c\u65b7\u8a5e\u8207\u8a5e\u6027\u6a19\u8a18\uff0c\u5c07 \u5728\u4e00\u4e9b\u6587\u672c\u5c0d\u4e2d\uff0c\u5176\u4e2d\u4e00\u53e5\u6709\u63d0\u4f9b\u6642\u9593\u8cc7\u8a0a\u3001\u4f46\u53e6\u4e00\u53e5\u6c92\u6709\uff0c\u53ef\u63a8\u8ad6\u9019\u5169\u53e5\u5982\u679c\u4e0d\u662f\u7368 \u7acb\u95dc\u4fc2\uff0c\u5c31\u662f\u6b63\u5411\u860a\u6db5\u95dc\u4fc2\u3002\u4ee5\u4e0b\u5217\u6587\u672c\u5c0d(S13,S14)\u70ba\u4f8b\uff0c\u5169\u53e5\u6709\u8a9e\u610f\u4e0a\u7684\u9ad8\u5ea6\u76f8\u95dc\uff0c \u4f46 S13 \u4e2d\u63d0\u5230\u4e86\u6642\u9593\u300c9 \u4e16\u7d00\u300d \uff0cS14 \u537b\u6c92\u6709\u63d0\u5230\u4efb\u4f55\u6642\u9593\uff0c\u56e0\u6b64\u8a72\u6587\u672c\u5c0d\u7684\u95dc\u4fc2\u6709\u53ef \u80fd\u662f\u6b63\u5411\u860a\u6db5\u3002 S21\uff1a\u7f8e\u806f\u793e\u548c\u4f0a\u666e\u7d22\u65af\u516c\u53f8\u6240\u9032\u884c\u7684\u6c11\u8abf\u986f\u793a\uff0c\u5e03\u5e0c\u7684\u65bd\u653f\u6eff\u610f\u5ea6\u5df2\u9996\u6b21\u6ed1\u843d\u5230 39% S22\uff1a\u4f0a\u666e\u7d22\u65af\u516c\u53f8\u548c\u7f8e\u806f\u793e\u6240\u9032\u884c\u7684\u6c11\u8abf\u986f\u793a\uff0c\u5e03\u5e0c\u7684\u65bd\u653f\u6eff\u610f\u5ea6\u5df2\u9996\u6b21\u6ed1\u843d\u5230 39% \u56e0\u6b64\u672c\u6587\u5b9a\u7fa9\u4e86\u7279\u5fb5\u300c\u8a5e\u5e8f\u4ea4\u63db\u300d \uff0c\u7c21\u7a31 WOE\uff0c\u5176\u503c\u5b9a\u7fa9\u5982\u4e0b: \u82e5\u6587\u672c\u5c0d\u7684\u7528\u8a5e\u5b8c\u5168\u4e00 \u6a23\u4f46\u662f\u8a5e\u5e8f\u4e0d\u4e00\u6a23\uff0c\u4e14\u9806\u5e8f\u4e0d\u540c\u7684\u8a5e\u4e26\u975e\u4ee5\u9023\u63a5\u8a5e\u9023\u63a5\uff0c\u5247\u8a72\u6587\u672c\u5c0d\u7684 WOE \u70ba 1\uff0c\u5426 Indicator Y N B F C I Macro-F1 F1 66.42 48.93 45.48 63.61 16.67 49.24 Decision tree[16] Precision 60.45 57.58 42.94 57.00 15.87 66.08 43.75 Recall 73.70 42.54 48.34 71.95 17.54 39.24 F1 69.80 60.37 55.78 65.72 6.61 56.16 \u7684\u8cc7\u6599\u90fd\u76f8\u7576\u7279\u5b9a\u3001\u6578\u91cf\u6709\u9650\uff0c\u56e0\u6b64\u9020\u6210\u77db\u76fe\u95dc\u4fc2\u7684\u66f4\u6df1\u5c64\u7279\u5fb5\u5c1a\u5f85\u767c\u6398\u3002\u53e6\u5916\uff0c\u672c\u6587 \u6240\u63d0\u65b9\u6cd5\u96d6\u7136\u4e5f\u5305\u542b\u8a9e\u610f\u7279\u5fb5\u985e\u5225\uff0c\u4f46\u90fd\u4ecd\u5c6c\u65bc\u8a9e\u610f\u7684\u9593\u63a5\u7279\u5fb5\uff0c\u4e26\u672a\u76f4\u63a5\u6e2c\u91cf\u8a9e\u610f\u3002 \u9019\u4e5f\u662f\u63a8\u8ad6\u6548\u80fd\u6709\u6240\u4fb7\u9650\u7684\u539f\u56e0\u3002 \u57fa\u65bc\u672c\u6587\u6240\u63d0\u65b9\u6cd5\uff0c\u672a\u4f86\u53ef\u9032\u4e00\u6b65\u63a2\u8a0e\u53ca\u7814\u7a76\u3002\u9996\u5148\uff0c\u672c\u6587\u63d0\u51fa\u7684 7 \u9805\u7279\u5fb5\u6709\u4e9b\u4ecd\u5f85\u9032 \u5176\u4e2d W k S9\uff1a\u99ac\u4f86\u897f\u4e9e\u539f\u70ba\u65e5\u672c\u96fb\u5b50\u696d\u8005\u773c\u4e2d\u6700\u4f73\u7684\u4e9e\u6d32\u6295\u8cc7\u6a19\u7684\uff0c\u73fe\u88ab\u4e2d\u570b\u5927\u9678\u53d6\u4ee3 S10\uff1a\u4e2d\u570b\u53d6\u4ee3\u7f8e\u570b\u6210\u70ba\u4e9e\u6d32\u7d93\u6fdf\u6838\u5fc3 S11\uff1a\u65e5\u672c\u662f\u6295\u8cc7\u99ac\u4f86\u897f\u4e9e\u7684\u4e09\u5927\u5916\u5546\u4e4b\u4e00 \u5247\u70ba-1\u3002 ID3 Precision 66.99 63.89 57.34 61.38 57.14 51.00 46.07 \u4e00\u6b65\u6539\u826f\uff0c\u4f8b\u5982\u4f7f\u7528\u540c\u7fa9\u8a5e\u3001\u5b58\u5728\u5426\u5b9a\u8a5e\u7b49\u7279\u5fb5\uff0c\u6240\u4f7f\u7528\u7684\u6e2c\u91cf\u65b9\u6cd5\u4ecd\u76f8\u7576\u7c21\u5316\u3002\u5982\u679c S13\uff1a\u97ed\u83dc\u65bc 9 \u4e16\u7d00\u50b3\u5165\u65e5\u672c Recall 72.86 57.21 54.30 70.73 3.51 62.50 \u80fd\u52a0\u4ee5\u6539\u826f\uff0c\u6548\u80fd\u61c9\u53ef\u6539\u5584\u3002\u53e6\u5916\uff0c\u8a9e\u6cd5\u7279\u5fb5\u5728\u76ee\u524d\u63d0\u51fa\u7684\u4e03\u9805\u7279\u5fb5\u4e2d\u50c5\u6709\u4e00\u9805\uff0c\u4f46\u5728 S14\uff1a\u97ed\u83dc\u66fe\u50b3\u5165\u65e5\u672c (\u4e94) \u9810\u6e2c\u6a21\u578b F1 72.78 50.72 61.67 63.48 10.94 49.30 \u5be6\u9a57\u904e\u7a0b\u4e2d\u767c\u73fe\u8a9e\u6cd5\u7279\u5fb5\u6709\u826f\u597d\u7684\u5340\u8fa8\u6548\u679c\u3002\u96d6\u7136\u76ee\u524d\u5df2\u7d93\u6709\u4e2d\u6587\u6587\u6cd5\u5256\u6790\u5de5\u5177\u63d0\u51fa\uff0c \u53e5\u5b50\u4ee5\u5b57\u5143\u8868\u73fe\u5f62\u5f0f\u8f49\u63db\u70ba\u8a5e\u5f59\u8868\u73fe\uff0c\u4e26\u6a19\u8a18\u6bcf\u500b\u8a5e\u5f59\u7684\u8a5e\u6027\u4ee5\u4fbf\u5f8c\u7e8c\u5206\u6790\u7279\u5fb5\u6642\u4f7f \u7528\u3002\u8a31\u591a\u4e2d\u6587\u65b7\u8a5e\u8207\u8a5e\u6027\u6a19\u8a18\u7cfb\u7d71\u5df2\u7d93\u88ab\u63d0\u51fa\uff0c\u4e5f\u6709\u5f88\u597d\u7684\u6b63\u78ba\u6027\u3002\u5728\u4e2d\u6587\u6587\u672c\u5c0d\u7684\u7814 S12\uff1a\u65e5\u672c\u6709\u6295\u8cc7\u99ac\u4f86\u897f\u4e9e \u672c\u6587\u5c07\u9019\u500b\u7279\u5fb5\u7a31\u70ba\u300c\u6642\u9593\u4e0d\u5c0d\u7a31\u300d \uff0c\u7c21\u7a31 OOT\u3002\u8a72\u7279\u5fb5\u5b9a\u7fa9\u5982\u4e0b\uff1a\u5982\u679c\u6587\u672c\u5c0d\u4e2d\u4e00\u53e5 \u6709\u6642\u9593\u8cc7\u8a0a\u800c\u53e6\u4e00\u53e5\u6c92\u6709\uff0c\u8a72\u7279\u5fb5\u503c\u70ba 1\uff0c\u53cd\u4e4b\u70ba-1\u3002 \u672c\u6587\u6240\u5b9a\u7fa9\u7684\u4e0a\u8ff0\u7279\u5fb5\u5c07\u6210\u70ba\u9810\u6e2c\u6a21\u578b\u7684\u8f38\u5165\u9805\u3002\u9810\u6e2c\u6a21\u578b\u5c07\u5229\u7528\u8a13\u7df4\u8cc7\u6599\u4e2d\u6bcf\u500b\u6587\u672c KC99-SVM Precision 62.96 70.67 53.11 55.03 50.00 58.29 46.35 Recall 86.42 39.55 73.51 75.00 6.14 42.71 \u4f46\u7528\u4ee5\u5206\u6790\u7279\u5fb5\u6642\u932f\u8aa4\u7387\u4ecd\u904e\u9ad8\u3002\u5982\u4f55\u4f7f\u7528\u6709\u6548\u7684\u6587\u6cd5\u5256\u6790\u5de5\u5177\u767c\u5c55\u8a9e\u6cd5\u7279\u5fb5\u53ef\u80fd\u662f\u80fd</td></tr></table>", | |
| "text": "\u5206\u5ee3\u6cdb\u3002\u4f8b\u5982\u77db\u76fe\u95dc\u4fc2\u5206\u6790\u53ef\u4ee5\u7528\u4f86\u81ea\u52d5\u5206\u6790\u4e0d\u540c\u77e5\u8b58\u4f86\u6e90\u6240\u63d0\u4f9b\u8cc7\u8a0a\u7684\u4e00\u81f4\u6027\u8207\u6b63\u78ba \u6027\uff0c\u9019\u5728\u7db2\u8def\u7684\u77e5\u8b58\u63d0\u4f9b\u8005\u5982 wiki \u4ee5\u53ca\u5229\u7528\u7db2\u8def\u8cc7\u6e90\u9032\u884c\u6578\u4f4d\u5b78\u7fd2\u7b49\u61c9\u7528\u90fd\u662f\u975e\u5e38\u91cd \u8981\u7684\u529f\u80fd\uff0c\u56e0\u70ba\u9019\u4e9b\u61c9\u7528\u5fc5\u9808\u78ba\u4fdd\u77e5\u8b58\u4f86\u6e90\u7684\u6b63\u78ba\u6027\u3002 \u76f8\u95dc\u554f\u984c\u7684\u8a31\u591a\u7814\u7a76\u5df2\u7d93\u88ab\u63d0\u51fa\uff0c\u70ba\u4e86\u8a55\u4f30\u9019\u4e9b\u65b9\u6cd5\u7684\u6548\u80fd\u5dee\u7570\uff0c\u6709\u8a31\u591a\u7684\u8a55\u4f30\u57fa\u6e96\u8cc7 \u6599\u88ab\u91cb\u51fa\uff0c\u4e5f\u8209\u8fa6\u4e86\u65b9\u6cd5\u6548\u80fd\u8a55\u6bd4\u3002\u5728 NTCIR-10 \u7684 RITE-2 \u4e2d\uff0c\u5404\u500b\u63a5\u53d7\u8a55\u6bd4\u7684\u65b9\u6cd5 \u5fc5\u9808\u5728\u7d66\u5b9a\u4e00\u500b\u7531\u5169\u500b\u53e5\u5b50\u6240\u7d44\u6210\u7684\u6587\u672c\u5c0d\u6642\uff0c\u80fd\u5920\u5206\u8fa8\u6587\u672c\u5c0d\u662f\u5c6c\u65bc\u96d9\u5411\u63a8\u8ad6\u95dc\u4fc2\u3001 \u6b63\u5411\u63a8\u8ad6\u95dc\u4fc2\u3001\u77db\u76fe\u95dc\u4fc2\u6291\u6216\u7368\u7acb\u95dc\u4fc2\u3002\u4ee5\u4e0b\u5217\u53e5\u5b50 S1 \u81f3 S5 \u6240\u7d44\u6210\u7684\u5404\u500b\u6587\u672c\u5c0d\u8209 \u4f8b\u8aaa\u660e\u3002 S2)\u800c\u8a00\uff0c\u53e5\u5b50 S2 \u4e2d\u7684\u6240\u6709\u8cc7\u8a0a\u5167\u5bb9\u53ef\u4ee5\u7531 S1 \u63a8\u8ad6\uff0c\u4f46 S1 \u4e2d\u6709\u4e00\u53e5\u5167\u5bb9 \u300c\u5e38\u898b\u7684\u852c\u83dc\u4e4b\u4e00\u300d\u662f\u4e0d\u80fd\u7531 S2 \u63a8\u65b7\u51fa\u4f86\u7684\uff0c\u56e0\u6b64\u9019\u6587\u53e5\u5c0d\u7a31\u70ba\u5177\u6709\u6b63\u5411\u95dc\u4fc2\u3002\u4ee5\u6587 \u672c\u5c0d(S2, S3)\u4f86\u8aaa\uff0c \u300c\u97ed\u83dc\u300d\u7684\u300c\u539f\u7522\u5730\u300d\u61c9\u8a72\u53ea\u6709\u4e00\u8655\uff0c\u4f46\u6b64\u4e8c\u53e5\u63cf\u8ff0\u97ed\u83dc\u7684\u539f\u7522\u5730\u662f \u4e0d\u540c\u7684\uff0c\u56e0\u6b64\u662f\u77db\u76fe\u7684\u8cc7\u8a0a\u3002\u5728\u63a8\u8ad6\u904e\u7a0b\u7522\u751f\u77db\u76fe\u7684\u6587\u672c\u5c0d\u7a31\u70ba\u5177\u6709\u77db\u76fe\u95dc\u4fc2\u3002\u800c\u8003\u616e \u6587\u672c\u5c0d(S3, S4)\uff0c\u96d6\u7136 S4 \u8207 S3 \u76f8\u4f3c\uff0c\u4f46\u8207\u6587\u672c\u5c0d(S2, S3)\u4e0d\u540c\uff0c \u300c\u97ed\u83dc\u300d\u8207\u300c\u6c34\u83dc\u300d\u7684 \u539f\u7522\u5730\u6c92\u6709\u4e92\u65a5\u95dc\u4fc2\uff0c\u56e0\u6b64\u9019\u5169\u53e5\u662f\u63cf\u8ff0\u4e0d\u540c\u7684\u4e8b\u4ef6\u800c\u4e0d\u662f\u540c\u4e00\u4e8b\u4ef6\u7684\u63cf\u8ff0\u4e0d\u540c\u3002\u9019\u6a23 \u7684\u60c5\u6cc1\u7a31\u8a72\u6587\u672c\u5c0d\u5177\u6709\u7368\u7acb\u95dc\u4fc2\u3002\u53c8\u8003\u616e\u6587\u672c\u5c0d(S4, S5)\uff0c\u5169\u53e5\u8868\u9054\u76f8\u540c\u7684\u77e5\u8b58\u5167\u5bb9\uff0c \u7531\u53e5\u5b50 S4 \u53ef\u4ee5\u63a8\u8ad6\u51fa S5\uff0c\u800c\u7531\u53e5\u5b50 S5 \u4e5f\u53ef\u63a8\u8ad6\u51fa S4\uff0c\u56e0\u6b64\u7a31\u6b64\u6587\u672c\u5c0d\u5177\u6709\u96d9\u5411\u63a8 \u8ad6\u95dc\u4fc2\u3002 Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing (ROCLING 2013) \u8868\u793a\u53e5\u5b50 S k \u4e2d\u6240\u6709\u8a5e\u7684\u96c6\u5408\uff0c|W k |\u8868\u793a\u96c6\u5408 W k \u4e2d\u7684\u8a5e\u7684\u6578\u91cf\u3002\u5c0d\u5169\u500b\u53e5\u5b50\u800c \u8a00\uff0c\u82e5 RWF \u8207 RWB \u5169\u8005\u540c\u6642\u90fd\u4f4e\uff0c\u986f\u793a\u5169\u53e5\u5b50\u7684\u8a9e\u610f\u53ef\u80fd\u76f8\u5dee\u904e\u5927\uff0c\u6b64\u6587\u672c\u5c0d\u5f88\u6709 \u53ef\u80fd\u662f\u7368\u7acb\u95dc\u4fc2\uff0c\u56e0\u70ba\u5169\u53e5\u8a71\u5305\u542b\u4e86\u4e0d\u540c\u7684\u5167\u5bb9\u3002\u8209\u4f8b\u4f86\u8aaa\uff0c\u4e0b\u5217\u6587\u672c\u5c0d(S9,S10)\u5169\u53e5 \u7684\u95dc\u4fc2\u70ba\u7368\u7acb\u95dc\u4fc2\uff0c\u5176 RWF \u70ba 0.05\u3001RWB \u70ba 0.16\u3002 Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing (ROCLING 2013) \u82e5\u5169\u53e5\u5b50\u7684 RWF \u548c RWB \u7684\u5dee\u503c\u5f88\u5927\uff0c\u5247\u8868\u793a\u6b64\u6587\u672c\u5c0d\u6709\u76f8\u8fd1\u7684\u8cc7\u8a0a\uff0c\u4f46\u5169\u53e5\u63d0\u4f9b\u7684 \u8cc7\u8a0a\u91cf\u4e00\u53e5\u8f03\u591a\u800c\u53e6\u4e00\u53e5\u8f03\u5c11\u3002\u9019\u662f\u6b64\u6587\u672c\u5c0d\u53ef\u80fd\u662f\u6b63\u5411\u860a\u6db5\u7684\u7dda\u7d22\u3002\u4f8b\u5982\u6587\u672c\u5c0d (S11,S12)\u70ba\u6b63\u5411\u860a\u6db5\uff0c\u5176 RWF \u70ba 0.30 \u800c RWB \u70ba 0.75\u3002\u57fa\u65bc\u4e0a\u8ff0\u89c0\u5bdf\uff0c\u672c\u6587\u5b9a\u7fa9\u4e86\u4e00 \u500b\u6587\u672c\u5c0d\u7279\u5fb5\u300c\u8a5e\u91cd\u758a\u7387\u5dee\u7570\u300d \uff0c\u7c21\u7a31 DRO\uff0c\u5b9a\u7fa9\u5982\u4e0b\uff1a \u5176\u4e2d TI \u548c TD \u662f\u5169\u500b\u9580\u6abb\u503c\u3002\u6839\u64da\u672c\u6587\u5be6\u9a57\u6240\u4f7f\u7528\u7684\u8a13\u7df4\u8cc7\u6599\uff0cTI \u548c TD \u7684\u503c\u5206\u5225\u70ba 0.6 \u8207 0.2\u3002 \u4ee3\u8868 S k \u53e5\u5b50\u4e2d\u6240\u6709\u8a5e\u6027\u7684\u96c6\u5408\uff0c|P k |\u70ba\u96c6\u5408 P k \u5167\u8a5e\u6027\u7684\u6578\u91cf\u3002\u82e5\u6587\u672c\u5c0d\u7684 RPF \u5920\u9ad8\uff0c\u4e14 RPF \u548c RPB \u7684\u5dee\u503c\u8d85\u904e\u95d5\u503c\uff0c\u5247\u9019\u500b\u6587\u672c\u5c0d\u5c31\u6709\u5f88\u9ad8\u7684\u6a5f\u7387\u662f\u6b63\u5411\u95dc\u4fc2\u3002\u56e0 \u6b64\u6587\u672c\u5c0d\u7684 DOP \u5b9a\u7fa9\u5982\u4e0b\uff1a \u5176\u4e2d TP \u548c TK \u662f\u5169\u500b\u9580\u6abb\u503c. \u6839\u64da\u672c\u6587\u5be6\u9a57\u6240\u4f7f\u7528\u7684\u8a13\u7df4\u8cc7\u6599, TP \u548c TK \u5404\u70ba 0.7 \u8207 Proceedings of the Twenty-Fifth Conference on Computational Linguistics and Speech Processing (ROCLING 2013) \u5c0d\u7684\u4e03\u9805\u7279\u5fb5\u503c\u8207\u5df2\u77e5\u7684\u6587\u672c\u5c0d\u860a\u6db5\u95dc\u4fc2\u4f5c\u70ba\u8a13\u7df4\u9810\u6e2c\u6a21\u578b\u4e4b\u7528\u3002\u5728\u8a13\u7df4\u5b8c\u6210\u5f8c\uff0c\u5c0d\u65bc \u8981\u9810\u6e2c\u7684\u6587\u672c\u5c0d\uff0c\u53ea\u8981\u8a08\u7b97\u8a72\u6587\u672c\u5c0d\u7684\u4e03\u9805\u7279\u5fb5\u503c\u5f8c\u8f38\u5165\u9810\u6e2c\u6a21\u578b\uff0c\u5373\u53ef\u5f97\u5230\u8a72\u6587\u672c\u5c0d \u95dc\u4fc2\u7684\u9810\u6e2c\u7d50\u679c\u3002\u672c\u6587\u4e3b\u8981\u76ee\u7684\u4e4b\u4e00\u5c31\u662f\u6bd4\u8f03\u6211\u5011\u5148\u524d\u63a1\u7528\u7684 decision tree \u4ee5\u53ca SVM \u65b9\u6cd5\u7684\u5dee\u7570\u3002SVM \u662f\u4e00\u500b\u76f8\u7576\u6210\u529f\u7684\u5206\u985e\u65b9\u6cd5\uff0c\u5df2\u7d93\u88ab\u5ee3\u6cdb\u61c9\u7528\u5728\u8a31\u591a\u9818\u57df\u7684\u7814\u7a76\u4e2d\u3002 \u672c\u6587\u5c07\u4ee5[15]\u6240\u63d0\u51fa\u7684 LibSVM \u4f5c\u70ba\u5be6\u4f5c SVM \u7684\u7cfb\u7d71\u3002\u53e6\u5916\u7531\u65bc\u5148\u524d\u63a1\u7528\u7684 decision tree \u662f\u7531\u5c08\u5bb6\u5efa\u7acb\uff0c\u672c\u6587\u4e5f\u5c07\u4ee5 ID3 \u6700\u4f73\u5316\u65b9\u6cd5\u81ea\u52d5\u67b6\u69cb decision tree \u4e26\u6bd4\u8f03\u6548\u80fd\u5dee\u7570\u3002 \u9805\u7279\u5fb5\u642d\u914d SVM \u7684\u65b9\u6cd5\u8207\u5176\u4ed6\u540c\u6a23\u4f7f\u7528 SVM \u4f46\u4e0d\u540c\u7279\u5fb5 \u7684\u65b9\u6cd5\u3002\u672c\u6587\u9078\u64c7\u5728 RITE2 \u4e2d\u4f7f\u7528 SVM \u8005\u6548\u80fd\u6700\u4f73\u7684 NTOUA[17]\u7cfb\u7d71\u4f5c\u70ba\u6bd4\u8f03\u5c0d\u8c61\u3002 \u5f9e\u8868\u4e8c\u53ef\u4ee5\u770b\u51fa\u672c\u6587\u7cfb\u7d71\u5728\u6574\u9ad4\u6548\u80fd\u4e0a\u8f03[17]\u70ba\u4f73\u3002\u7531\u65bc\u8a72\u7cfb\u7d71\u4f7f\u7528 20 \u7a2e\u7279\u5fb5\u800c\u672c\u6587 \u50c5\u4f7f\u7528 7 \u7a2e\u7279\u5fb5\uff0c\u56e0\u6b64\u53ef\u77e5\u672c\u6587\u6240\u63d0\u65b9\u6cd5\u80fd\u4ee5\u8f03\u5c11\u6578\u91cf\u7684\u7279\u5fb5\u9054\u5230\u66f4\u597d\u7684\u6548\u80fd\u3002 \u8868\u4e8c\u3001\u4ee5 SVM \u9032\u884c\u5206\u985e\u4f46\u63a1\u7528\u4e0d\u540c\u7279\u5fb5\u7684\u65b9\u6cd5\u9593\u6548\u80fd\u6bd4\u8f03 Precision 68.64 66.48 53.06 53.99 36.25 52.73 46.32 Recall 74.95 59.20 51.66 80.49 25.44 30.21 F1 72.78 50.72 61.67 63.48 10.94 49.30 KC99-SVM Precision 62.96 70.67 53.11 55.03 50.00 58.29 46.35 Recall 86.42 39.55 73.51 75.00 6.14 42.71", | |
| "type_str": "table", | |
| "html": null | |
| } | |
| } | |
| } | |
| } |