模糊限制信息检测中融合方法的研究
|
摘要
作为生物信息抽取的一个重要环节,生物医学领域的模糊限制信息检测旨在区分生物医学文献中的模糊限制信息与事实信息,避免将模糊限制信息作为事实信息用于信息抽取。
近年来,随着大规模模糊限制信息语料库的构建,虽然模糊限制信息检测研究已经取得了一定的进展,但是模糊限制信息范围检测性能尚未达到60%,距离实用化还有一段距离。这是由于模糊限制信息范围检测任务比较复杂,具有依赖于语义和句法结构的特点,单纯基于一个统计模型难以满足模糊限制信息范围检测这个复杂任务的处理需求。融合方法可以将自然语言处理任务中的多类特征、多种方法、多个模型有效结合起来,避免了单一模型的片面性,实现准确、健壮的自然语言处理。
本文针对模糊限制信息检测中的融合方法展开研究,内容主要包括：
(1)研究基于复合核函数融合结构化特征与平面特征的模糊限制信息范围检测。
重点研究了基于短语的模糊限制信息范围的结构化表达形式,利用卷积树核函数捕获模糊限制信息范围的结构化信息,减小结构化信息平面化时所引起的信息丢失。然后将基于结构化特征的卷积树核函数与基于平面特征的多项式核函数通过复合核函数集成起来。得到的复合核函数取得了比单独使用两种核函数都好的检测性能。
(2)研究统计方法和规则方法相结合的模糊限制信息范围检测。
通过统计方法和规则方法的结合,融合基于短语结构和基于依存结构的模糊限制信息范围检测系统。首先分别利用短语结构建立基于支持向量机(Support Vector Machine,SVM)的模糊限制信息范围检测子系统,利用依存结构构建基于规则的模糊限制信息范围检测子系统。然后将两个子系统的检测结果作为两个独立的特征,引入条件随机域(Conditional Random Field, CRF)模型进行融合。这种融合方法有效地利用了短语结构和依存结构,实现了统计方法和规则方法的结合,以及SVM机器学习方法和CRF机器学习方法的结合。统计和规则相结合的模糊限制信息范围检测方法取得了比单独使用两种方法都好的检测结果。
(3)研究多分类器相融合的模糊限制信息范围检测。
提出一种基于投票策略的模糊限制信息范围检测方法,首先分别基于SVM、CRF、最大间隔马尔可夫网络(Max-Margin Markov Networks,M3N)、以及本文的统计和规则结合的方法,以前向和后向两个解析方向构建八个基本分类器,再分别采用多数投票、分类器加权投票和词性加权投票三种投票策略融合八个基本分类器的结果。基于投票策略的模糊限制范围检测系统都取得了稳定的且比其中最优分类器更好的分类性能。
本文的主要成果在于对模糊限制信息检测中的融合策略进行了深入研究,探索了模糊限制信息检测任务中平面特征与结构特征的融合、基于统计方法与基于规则方法的融合、多分类器的融合。提出了基于复合核函数的模糊限制信息范围检测方法,实现了模糊限制信息检测中结构化特征与平面特征的融合；提出了基于统计方法和规则方法的结合,有效利用短语结构和依存结构的模糊限制信息范围检测方法；提出了基于投票策略的多分类器模糊限制信息范围检测方法。这些研究有效地提高了生物医学领域模糊限制信息检测性能,对今后自然语言处理中融合策略的研究提供了有益的借鉴。
To distinguish factual and uncertain information in biological texts, hedged information detection is extremely important for biomedical information extraction, which avoids extracting speculative information as factual information.
As large-scale tagged Bioscope corpus has become available these days, studies in detecting hedge scope have been developed.However, the performance for hedge scope detection is still less than60%.There is a considerable gap between academic researches and practical applications.Hedge scope detection is rather complicated as it falls within the scope of semantic analysis of sentences exploiting syntactic patterns. For complicated hedge scope detection task, there exists no reliable and simple way to achieve a satisfactory performance.Every kinds of feature, every method, every model has its advantages and limitations, and they are complement for each other. So how to combine the advantages of various kinds of features, methods, models, and avoid one-sidedness of a single model to develop high-accurate fusion hedge detection systems, become an important theme of natural language processing.
This paper focuses on the fusion methods for hedge detection. The main works are listed as follows:
1.The approach to hedge scope detection using a composite kernel which combines structured and flat features.Four phrase-based structured features over a parse tree are explored for hedge scope learning to capture the critical syntactic structure by the convolution tree kernel.The convolution tree kernel that exploits the syntactic structured features and the polynomial kernel that exploits the flat features are combined into a composite kernel.The composite kernel outperforms either of the two individual kernels.
2.The hybrid approach based on rules and statistics to hedge scope detection, which can also combine phrase structures and dependency structures.First, phrase structures and dependency structures are used for hedge scope detection respectively.Phrase structures are adapted as important features for hedge scope detection by a Support Vector Machine (SVM)-based model.Dependency structures are used to detect hedge scope by a rule-based method. Then, the phrase-based system and the dependency-based system are combined by a Conditional Random Field (CRF)-based model, which simply extends the feature vectors with the scope tags generated by the two individual phrase-based and dependency-based systems. The combination of rule-based and statistics-based approaches,the combination of phrase structures and dependency structures,and the combination of SVM and CRF in our fusion system are all factors for effective scope detection. Experimental results show that phrase structures and dependency structures are both effective for hedge scope detection and their combination can improve the scope detection performance further.
3.The voting technique for detecting hedge scope.First we construct eight classifiers based on CRF,SVM, Max-Margin Markov Network (M3N) and our rule-based and statistics-based combination approach, time two directions (forward and backward).Then three different voting schemes:(1)majority voting;(2) weighted voting by the accuracy of the component classifier;(3)POS weighted voting by the accuracy of the component classifier on all tokens which have the same POS,are adapted to voting-based hedge scope detection. The experimental results show that voting may result in improvement over their component classifiers by combining their individual advantages.
This paper explores the fusion methods to hedge detection, including the combination of structured and flat features,the hybrid approach based on rule-based and statistics-based approaches, the method of multiple classifier fusion.The major contributions of this paper lie on the proposal of a phrase-based approach to hedge scope detection using a composite kernel which combines structured and flat features;the proposal of the hybrid approach based on rules and statistics to hedge scope detection, which can also combine phrase structures and dependency structures;the proposal of the voting scheme to detect hedge scope which combines many individual classifiers to exploit the unique advantage of each classifier. This work improves the hedge detection performance significantly, and exhibits reference value to the future research in fusion methods to natural language processing.
近年来,随着大规模模糊限制信息语料库的构建,虽然模糊限制信息检测研究已经取得了一定的进展,但是模糊限制信息范围检测性能尚未达到60%,距离实用化还有一段距离。这是由于模糊限制信息范围检测任务比较复杂,具有依赖于语义和句法结构的特点,单纯基于一个统计模型难以满足模糊限制信息范围检测这个复杂任务的处理需求。融合方法可以将自然语言处理任务中的多类特征、多种方法、多个模型有效结合起来,避免了单一模型的片面性,实现准确、健壮的自然语言处理。
本文针对模糊限制信息检测中的融合方法展开研究,内容主要包括：
(1)研究基于复合核函数融合结构化特征与平面特征的模糊限制信息范围检测。
重点研究了基于短语的模糊限制信息范围的结构化表达形式,利用卷积树核函数捕获模糊限制信息范围的结构化信息,减小结构化信息平面化时所引起的信息丢失。然后将基于结构化特征的卷积树核函数与基于平面特征的多项式核函数通过复合核函数集成起来。得到的复合核函数取得了比单独使用两种核函数都好的检测性能。
(2)研究统计方法和规则方法相结合的模糊限制信息范围检测。
通过统计方法和规则方法的结合,融合基于短语结构和基于依存结构的模糊限制信息范围检测系统。首先分别利用短语结构建立基于支持向量机(Support Vector Machine,SVM)的模糊限制信息范围检测子系统,利用依存结构构建基于规则的模糊限制信息范围检测子系统。然后将两个子系统的检测结果作为两个独立的特征,引入条件随机域(Conditional Random Field, CRF)模型进行融合。这种融合方法有效地利用了短语结构和依存结构,实现了统计方法和规则方法的结合,以及SVM机器学习方法和CRF机器学习方法的结合。统计和规则相结合的模糊限制信息范围检测方法取得了比单独使用两种方法都好的检测结果。
(3)研究多分类器相融合的模糊限制信息范围检测。
提出一种基于投票策略的模糊限制信息范围检测方法,首先分别基于SVM、CRF、最大间隔马尔可夫网络(Max-Margin Markov Networks,M3N)、以及本文的统计和规则结合的方法,以前向和后向两个解析方向构建八个基本分类器,再分别采用多数投票、分类器加权投票和词性加权投票三种投票策略融合八个基本分类器的结果。基于投票策略的模糊限制范围检测系统都取得了稳定的且比其中最优分类器更好的分类性能。
本文的主要成果在于对模糊限制信息检测中的融合策略进行了深入研究,探索了模糊限制信息检测任务中平面特征与结构特征的融合、基于统计方法与基于规则方法的融合、多分类器的融合。提出了基于复合核函数的模糊限制信息范围检测方法,实现了模糊限制信息检测中结构化特征与平面特征的融合；提出了基于统计方法和规则方法的结合,有效利用短语结构和依存结构的模糊限制信息范围检测方法；提出了基于投票策略的多分类器模糊限制信息范围检测方法。这些研究有效地提高了生物医学领域模糊限制信息检测性能,对今后自然语言处理中融合策略的研究提供了有益的借鉴。
To distinguish factual and uncertain information in biological texts, hedged information detection is extremely important for biomedical information extraction, which avoids extracting speculative information as factual information.
As large-scale tagged Bioscope corpus has become available these days, studies in detecting hedge scope have been developed.However, the performance for hedge scope detection is still less than60%.There is a considerable gap between academic researches and practical applications.Hedge scope detection is rather complicated as it falls within the scope of semantic analysis of sentences exploiting syntactic patterns. For complicated hedge scope detection task, there exists no reliable and simple way to achieve a satisfactory performance.Every kinds of feature, every method, every model has its advantages and limitations, and they are complement for each other. So how to combine the advantages of various kinds of features, methods, models, and avoid one-sidedness of a single model to develop high-accurate fusion hedge detection systems, become an important theme of natural language processing.
This paper focuses on the fusion methods for hedge detection. The main works are listed as follows:
1.The approach to hedge scope detection using a composite kernel which combines structured and flat features.Four phrase-based structured features over a parse tree are explored for hedge scope learning to capture the critical syntactic structure by the convolution tree kernel.The convolution tree kernel that exploits the syntactic structured features and the polynomial kernel that exploits the flat features are combined into a composite kernel.The composite kernel outperforms either of the two individual kernels.
2.The hybrid approach based on rules and statistics to hedge scope detection, which can also combine phrase structures and dependency structures.First, phrase structures and dependency structures are used for hedge scope detection respectively.Phrase structures are adapted as important features for hedge scope detection by a Support Vector Machine (SVM)-based model.Dependency structures are used to detect hedge scope by a rule-based method. Then, the phrase-based system and the dependency-based system are combined by a Conditional Random Field (CRF)-based model, which simply extends the feature vectors with the scope tags generated by the two individual phrase-based and dependency-based systems. The combination of rule-based and statistics-based approaches,the combination of phrase structures and dependency structures,and the combination of SVM and CRF in our fusion system are all factors for effective scope detection. Experimental results show that phrase structures and dependency structures are both effective for hedge scope detection and their combination can improve the scope detection performance further.
3.The voting technique for detecting hedge scope.First we construct eight classifiers based on CRF,SVM, Max-Margin Markov Network (M3N) and our rule-based and statistics-based combination approach, time two directions (forward and backward).Then three different voting schemes:(1)majority voting;(2) weighted voting by the accuracy of the component classifier;(3)POS weighted voting by the accuracy of the component classifier on all tokens which have the same POS,are adapted to voting-based hedge scope detection. The experimental results show that voting may result in improvement over their component classifiers by combining their individual advantages.
This paper explores the fusion methods to hedge detection, including the combination of structured and flat features,the hybrid approach based on rule-based and statistics-based approaches, the method of multiple classifier fusion.The major contributions of this paper lie on the proposal of a phrase-based approach to hedge scope detection using a composite kernel which combines structured and flat features;the proposal of the hybrid approach based on rules and statistics to hedge scope detection, which can also combine phrase structures and dependency structures;the proposal of the voting scheme to detect hedge scope which combines many individual classifiers to exploit the unique advantage of each classifier. This work improves the hedge detection performance significantly, and exhibits reference value to the future research in fusion methods to natural language processing.
引文
[1]冯志伟.计算语言学基础[M].北京：商务印书馆,2001.
[2]Manning C D, Schutze H. Foundations of statistical natural language processing[M],MIT Press,1999.
[3]Daelemans W, Buchholz S, Veenstra J. Memory-based shallow parsing[C].In Proceedings of CoNLL-1999,1999:53-60.
[4]Roth D. Memory based learning in NLP[R].Technical Report 2125,Urbana, Illinois,1999.
[5]Zavrel J, Daelemans W. Memory-based learning:using similarity for smoothing[C]. Proceedings of the Thirty-Fifth Annual Meeting of the Association for Computational Linguistics and Eighth Conference of the European Chapter of the Association for Computational Linguistics,1997:436-443.
[6]Quinlan J R. C4.5:Programs for machine learning[M]. Morgan Kaufmann Publishers,1993.
[7]Heeman P A. POS tags and decision trees for language modeling[C]. SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora, College Park, Maryland, June 1999:129-137.
[8]Ratnaparkhi A. A maximum entropy model for part-of-speech tagging[C]. The 1996 Conference on Empirical Methods on Natural Language Processing,1996:133-142.
[9]Ratnaparkhi A. A liner observed time statistical parser based on maximum entropy models[C]. The 1997 Conference on Empirical Methods on Natural Language Processing, 1997:1-10.
[10]Cortes C, Vapnik V N. Support vector networks. Machine Learning[J],1995,20:273-297.
[11]Vapnik V N. Statistical learning theory[M]. Wiley-Interscience,1998.
[12]Lafferty J, McCallum A, Pereira F. Conditional random fields:Probabilistic models for segmenting and labelling sequence data[C]. Proceedings of the Eighteenth International Conference on Machine Learning,2001:282-289.
[13]Taskar B, Guestrin C, Koller D. Max-Margin markov networks[C]. Proceeding of Neural Information Processing Systems Conference,2003.
[14]Lakoff G. Hedges:a study in meaning criteria and the logic of fuzzy concepts[J]. Journal of Philosophical Logic,1973,2(4):458-508.
[15]Light M, Qiu X Y, Srinivasan P. The language of bioscience:facts, speculations, and statements in between[C]. Proceedings of the BioLINK,2004:17-24.
[16]Vincze V, Szarvas G, Farkas R, Mora G, Csirik J. The BioScope corpus:biomedical texts annotated for uncertainty, negation and their scopes[J]. BMC Bioinformatics, 2008,9(11):S9.
[17]Szarvas G. Hedge classification in biomedical texts with a weakly supervised selection of keywords[C]. Proceedings of Association for Computational Linguistics, HLT 2008: 281-289.
[18]Medlock B, Briscoe T. Weakly supervised learning for hedge classification in scientific literature [C]. The 45th Annual Meeting of the Association of Computational Linguistics, 2007:992-999.
[19]Farkas R, Vincze V, Mora G, Csirik J, Szarvas G. The CoNLL 2010 shared task:learning to detect hedges and their scope in natural language text[C]. Proceedings of the CoNLL2010 Shared Task,2010:1-12.
[20]Hyland K. Writing without conviction hedging in science research articles[J]. Applied Linguistics,1996,7:433-454.
[21]Meyers G. The pragmatics of politeness in scientific [J]. English for Specific Purposes, 2000,19(13):175-187.
[22]Meyers F. Procrustes'recipe:hedging and positivism[J]. English for Specific Purposes: An International Journal 2000,19(2):175-189.
[23]Skelton J. The representation of truth in academic medical writing[J]. Applied Linguistics,1997,18(2):121-140.
[24]Varttala, T. Remarks on the communicative functions of hedging in popular scientific and specialist research articles on medicine [J]. English for Specific Purposes,1999, 18(2):177-200.
[25]赵英玲.英语科技语体中的模糊限制语[J].外语与外语教学,1999,124(9)：15-17.
[26]冯茵,周榕.学术论文摘要中模糊限制语的调查与分析—基于英语专业毕业论文与国外期刊论文的对比研究[J].外国语言文学,2007,92(2)：108-112.
[27]Prince, E F, Frader J, Bosk C, Dipietro R J. On the hedging in physician-physician discourse[C]. In Linguistics and the Professions. J. Robert (ed.) Norwood:Ablex, 1982:83-97.
[28]喻亚男,徐畅贤.论英语模糊限制语的语用功能[J].益阳师专学报,2002,23(2)：102-103.
[29]范晓晖.医患会话中模糊限制语的语用功能[J].西北医学教育,2006,14(6)：740-742.
[30]范晓晖,孙圆圆.医学论文英文摘要模糊限制语的对比研究[J]. 西北医学教育,2008,16(6)：1187-1189.
[31]王舟.英汉学术论文摘要中模糊限制语的对比研究[J].华中科技大学学报,2006,22(6)：59-63.
[32]Kilicoglu H, Bergler S. Recognizing speculative language in biomedical research articles:a linguistically motivated perspective[C]. Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing, Columbus, Ohio, June, Association for Computational Linguistics,2008:46-53.
[33]Ganter V, Strube M. Finding hedges by chasing weasels:Hedge detection using Wikipedia tags and shallow linguistic features[C]. Proceedings of the ACL-IJCNLP 2009 Conference Short Papers, Suntec, Singapore, August, Association for Computational Linguistics, 2009:173-176.
[34]Zhou H W, Li X Y, Huang D G, Li Z Z, and Yang Y S. Exploiting multi-features to detect hedges and their scope in biomedical texts [C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:56-63.
[35]Li X X, Shen J P, Gao X, Wang X. Exploiting rich features for detecting hedges and their scope[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:36-41.
[36]Zhang S D, Zhao H, Zhou G D, Lu B L. Hedge detection and scope finding by sequence labeling with normalized feature selection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[37]Ji F, Qiu X P, Huang X J, Detecting hedge cues and their scopes with average perceptron[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July. Association for Computational Linguistics,2010:139-146.
[38]Lin Chen, Barbara Di Eugenio.A lucene and maximum entropy model based hedge detection system[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July. Association for Computational Linguistics,2010:114-119.
[39]Prabhakaran V. Uncertainty learning using SVMs and CRFs[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:132-137.
[40]Zhao Q., Sun C. J., Liu B. Q., Cheng Y.. Learning to detect hedges and their scope using CRF[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:100-105.
[41]王正群.手写体汉字识别研究[D].南京：南京理工大学,2001.
[42]Tang B Z, Wang X L, Wang X, Yuan B, Fan S X. A cascade method for detecting hedges and their scope in natural language text[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:25-29.
[43]Ozgur A, Radev D R. Detecting speculations and their scopes in scientific text[C]. Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, Singapore, August,Association for Computational Linguistics.2009:1398-1407.
[44]Velldal E,Ovrelid L, Oepen S. Resolving speculation:MaxEnt cue classification and dependency-based scope rules[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:48-55.
[45]Morante R, Asch V V, Daelemans W. Memory-based resolution of insentence scopes of hedge cues[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:48-55.
[46]Kilicoglu H, Bergler S. A high-precision approach to detecting hedges and their scopes[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[47]Zhang S D, Zhao H, Zhou G D, Lu B L. Hedge detection and scope finding by sequence labeling with procedural feature selection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[48]Zhu Q M, Li J H, Wang H L, Zhou G D. A unified framework for scope learning via simplified shallow semantic parsing[C]. Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing,2010:714-724.
[49]Morante R, Daelemans W. Learning the scope of hedge cues in biomedical texts[C]. Proceedings of the BioNLP 2009 Workshop, pages, Boulder, Colorado, June, Association for Computational Linguistics,2009:28-36.
[50]Rei M, Briscoe T. Combining manual rules and supervised learning for hedge cue and scope detection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:56-63.
[51]Joachims T. Text categorization with support vector machines:learning with many relevant features[C]. Proceedings 10th European Conference on Machine Learning(ECML), Springer Verlag,1998.
[52]Dumais S, Platt J, Heckerman D, Sahami M. Inductive learning algorithms and representations for text categorization[R]. Technical Report, Microsoft Research,1998:148-152.
[53]Lodhi H, Saunders C, Sahwe-Taylor J, Cristianini N, Watkins C. Text classification using string kernels[C]. In Neural Information Processing Systems (NIPS), MIT Press, 2001:563-569.
[54]Mitra V, Wang C J, Banerjee S. Text classification:A least square support vector machine approach[J]. Applied Soft Computing,2007,7(3):908-914.
[55]Lee K S, Kageura K. Virtual relevant documents in text categorization with support vector machines [J]. Information Processing and Management:an International Journal, 2007,43(4):902-913.
[56]Balcazar J L, Dai Yang, Watanabe O. Provably fast training algorithms for support vector machines[C]. Proceedings of IEEE International Conference on Data Mining,2001,43-50.
[57]Mangasarian O L. A finite Newton method for classification[J]. Optimization Methods and Software,2002,17(5):913-929.
[58]Keerthi S, DeCoste D. A modified finite Newton method for fast solution of large scale linear SVMs[J]. Journal of Machine Learning Research.2005,6(5):341-361.
[59]Dong J X, Krzyzak A, Suen C Y. An improved handwritten Chinese character recognition system using support vector machine[J]. Pattern Recognition Letters.2005, 26 (12):1849-1856.
[60]Dong J X, Krzyzak A, Suen C Y. Fast SVM training algorithm with decomposition on very large data sets[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(4):603-618.
[61]Wu Y C, Fan T K, Lee Y S, et al. Extracting named entities using support vector machines[J]. Lecture Notes in Bioinformatics(LNBI):Knowledge Discovery in Life Science Literature,2006,3886:91-103.
[62]Kuto T, Matsumoto Y. Chunking with support vector machines[C]. Proceeding of NAACL, 2001:192-199.
[63]Kuto T, Matsumoto Y. Japanese dependency analysis based on support vector machines[C]. In Empirical Methods in Natural Language processing and Very Large Corpora,2000:18-25.
[64]Scholkopf B. Input space versus feature dpace in kernel-based methods[J]. IEEE Translations on Neural Networks,1999,10(5):1000-1016.
[65]Muller K R. An introduction to kernel-based learning algorithms [J]. IEEE Translations on Neural Networks,2001,12(2):181-201.
[66]Boser B E, Guyon L M, and Vapnik V N. A training algorithm for optimal margin classifiers[C]. Proceedings of the 5th annual ACM workshop on computational learning theory. ACM Press,1992:144-152.
[67]Osuna E, Freund R, Girosi F. Training support vector machines:an application to face detection[C]. Proceedings CVPR'97,1997:130-136.
[68]Joachims T. Making large-scale SVM learning practical[C]. Conference on Advances in Kernel Methods Support Vector Learning, MIT Press,1999:169-184.
[69]Platt J C, Scholkopf B, Burges C, Smola A. Fast training of support vector machines using sequential minimal optimization[C]. Conference on Advances in Kernel Methods:Support Vector Learning, MIT Press,1998:185-208.
[70]Platt J C. Using analytic QP and sparseness to speed training of support vector machines[C]. Conference on Advances in neural information processing systems Ⅱ,1999:557-563.
[71]Eddy S R. Hidden markov models[J]. Current Opinion in Structural Biology,1996, 6(3):361-365.
[72]McCallum A, Freitag D, Pereira F. Maximum entropy markov models for information extraction and segmentation[C]. Proceedings of the Seventeenth International Conference on Machine Learning,2000:591-598.
[73]Kudo T, Yamamoto K, Matsumoto Y. Applying conditional random fields to Japanese morphological analysis[C]. Proceedings of EMNLP 2004,2004:230-237.
[74]黄德根,焦世斗,周惠巍.基于子词的双层CRFs中文分词[J].计算机研究与发展,2010,47(5)：962-968.
[75]周俊生,戴新宇,尹存燕,陈家骏.基于层叠条件随机场模型的中文机构名自动识别[J].电子学报,2006,34(5)：804-809.
[76]McCallum A and Li W. Early results for named entity recognition with conditional random fields, feature induction and web-enhanced lexicons[C]. Proceeding of CoNLL 2003, 2003:188-191.
[77]Sha F, Pereira F. Shallow parsing with conditional random fields[C]. Proceedings of Human Language Technology-NAACL,2003:134-141.
[78]Forney J G. The viterbi algorithm[J]. Proceedings of the IEEE,1973,61(3):268-278.
[79]Johansen S, Juselius K. Maximum likelihood estimation and inference on cointegration-with applications to the demand for money[J].Oxford Bulletin of Economics and Statistics,1990,52(2):169-210.
[80]Darroch J N, Ratcliff D. Generalized iterative scaling for log-linear models[J].The Annals of Mathematical Statistics,1972,43(5):1470-1480.
[81]Malouf R. A comparison of algorithms for maximum entropy parameter estimation[C]. Proceeding of the 6th conference on Natural language learning,2002:1-7.
[82]Shewchuk J R. An introduction to the conjugate gradient method without the agonizing pain[M]. Edition 1 1/4, School of Computer Science, Carnegie Mellon University,1994.
[83]Byrd R H, Nocedal J, Schnabel R B. Representation of quasi-Newton matrices and their use in limited memory methods[J]. Mathematical Programming,1994,63:129-156.
[84]Sha F, Pereira F. Shallow parsing with conditional random fields[C]. Proceedings of Humna Language Technology NAACL, Edmonton, Canada,2003:213-220.
[85]Collins M, Duffy N. Convolution kernels for natural language[C]. Conference on Advances in Neural Information Processing Systems,2001:625-632.
[86]Kong F, Zhou G D, Qian L H, Zhu Q M. Dependency-driven anaphoricity determination for coreference resolution[C]. Proceedings of the 23rd International Conf.on Computational Linguistics,2010:599-607.
[87]Zhou G D, Kong F. Global learning of noun phrase anaphoricity in coreference resolution via label propagation[C].Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing,2009:978-986.
[88]Zhang M, Zhang J, Su J. Exploring syntactic features for relation extraction using a convolution tree kernel[C].Proceedings of the Human Language Technology Conf.of the North American Chapter of the ACL,2006:288-295.
[89]Moschitti A, Pighin D, Basili R. Semantic role labeling via tree kernel joint inference[C]. Proceedings of the 10th Conference on Computational Natural Language Learning, New York, USA,2006:61-68.
[90]Tsuruoka Y, Tateishi Y, Kim J D, Ohta T, McNaught J, Ananiadou S, Tsujii J. Developing a robust part-of-speech tagger for biomedical text[C]. Conference on Advances in Informatics 2005:382-392.
[91]Christiane F. WordNet and wordnets[M]. Encyclopedia of Language and Linguistics, Second Edition, Oxford:Elsevier,2005:665-670.
[92]Liu Y, Tan Q, Shen K, The word segmentation rules and automatic word segmentation methods for Chinese information processing[M]. QingHua University Press and GuangXi Science and Technology Press,1994:2-36.
[93]Petrov S, Barrett L, Thibaux R, Klein D. Learning accurate, compact, and interpretable tree annotation[C]. Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics,2006:433-440.
[94]Petrov S, Klein D. Improved inference for unlexicalized parsing[C]. Proceedings of NAACL HLT 2007:404-411.
[95]Zhou G D, Zhang M, Ji D H, and Zhu Q M. Tree kernel-based relation extraction with context-sensitive structured parse tree information[C]. Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational,2007:728-736.
[96]Joachims T, Cristianini N, Shawe-Taylor J. Composite kernels for hypertext categorization[C]. Proceedings of ICML-2001,2001:250-257.
[97]Zhang M, Zhang J, Su J, and Zhou G D. A composite kernel to extract relations between entities with both flat and structured features[C]. ACL 44th Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the Association for Computational Linguistics,2006:825-832.
[98]车万翔.基于核方法的语义角色标注[D].哈尔滨：哈尔滨工业大学,2008.
[99]钱龙华.命名实体间语义关系抽取研究[D].苏州：苏州大学,2009.
[100]Moschitti A. A study on convolution kernels for shallow semantic parsing[C]. Proceedings of the 42th Conference on Association for Computational Linguistic, Barcelona, Spain,2004:335-342.
[101]Zhou H W, Huang D G, Li X Y, Yang Y S. A staged and distributed strategy for bio-entity recognition[J]. International Journal of Information,14(10):3527-3536.
[102]冯志伟.自然语言处理的形式模型[M].安徽：中国科学技术大学出版社,2010.
[103]Earley J. An efficient context-free parsing algorithm[J]. Communications of the association for computing machinery,1970,13(2):94-102.
[104]宗成庆.统计自然语言处理[M].北京：清华大学出版社,2008.
[105]赵伟,戴新宇,尹存燕,陈家骏.一种规则与统计相结合的汉语分词方法[J].计算机工程与科学,2003(3)：23-25
[106]姜尚仆,陈群秀.基于规则和统计的日语分词和词性标注的研究[J].中文信息学报,2010,24(1)：117-122.
[107]周强.规则与统计结合的汉语词类标注方法[J].中文信息学报,1995,9(2)：1-10.
[108]姜柄圭,张秦龙,谌贻荣,常宝宝.面向机器辅助翻译的汉语语块自动抽取研究[J].中文信息学报,2007,21(1)：9-16.
[109]李素建,刘群,白硕.统计和规则相结合的汉语组块分析[J].计算机研究与发展,2002,39(4)：385-391.
[110]苗艳军;李军辉;周国栋.统计和规则相结合的并列结构自动识别[J].计算机应用研究,2009,26(9)：3403-3406.
[111]傅间莲,陈群秀.基于规则和统计的中文自动文摘系统[J]. 中文信息学报,2006,20(5)：10-16.
[112]昝红英,左维松,张坤丽,吴云芳.规则和统计相结合的情感分析研究[J].计算机工程与科学,2011,33(5),146-150.
[113]郭宏蕾,胡岗.统计与规则相结合的目标语言生成策略[C].2002年全国机器翻译研讨会,002：110-115.
[114]Eisele A, Federmann C, Uszkoreit H, Saint-Amand H, Kay M, Jellinghaus M, Hunsicker S, Herrmann T, Chen Y. Hybrid machine translation architectures within and beyond the EuroM atrix project[C]. Proceedings of the 12th annual conference of the European Association for Machine Translation (EAMT 2008). Hamburg, Germany,2008:27-34.
[115]徐金安.理性主义与经验主义相结合的机器翻译研究策略[J].计算机科学,2011,38(6)：223-229.
[116]Sagae K, Tsujii J. Dependency parsing and domain adaptation with LR models and parser ensembles. Proceedings of the CoNLL 2007 Shared Task[C]. Joint Conferences on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL'07).2007:1044-1050.
[117]Spengler A. Maximum margin markov networks for XML tag relabeling[D]. Karlsruhe: University of Karlsruhe,2005.
[118]Kleinberg E M. A Mathematically rigorous foundation for supervised learning[J]. Lecture Notes in Computer Science,2000,1857:67-76.
[119]Allein E L, Schapire R E, and Singer Y. Reducing multiclass to binary:a unifying approach for margin classifiers[J]. Journal of Machine Learning Research, 2000,1(9):113-141
[120]周俊生,戴新宇,尹存燕,陈家俊.自然语言信息抽取中的机器学习方法研究[J].计算机科学,2005,32(3)：186-189.
[121]Halteren H V, Zavrel J, Daelemans W. Improving accuracy in word class tagging through the combination of machine learning systems[J]. Coynputatioual Linguistics, 2001,27(2):199-230.
[122]Rahman A F R, Fairhurst M C. Multiple calssifier decision combination strategies for character recognition:a review[J]. International Journal on Document Analysis and Recognition.2003,5(4):166-194
[123]Raudys S, Roli F. The behavior knowledge space fusion method:analysis of generalization error and strategies for performance improvement [C]. Proceedings of 4th International Workshop on Multiple Classifier Systems (MCS). Lecture Notes in Computer Science 2709,2003:55-64.
[124]Lam L, Suen C Y. Application of majority voting to pattern recognition:an analysis of its behavior and performance[J]. IEEE Transactions on systems, man, and Cybernetics-Part A:Systems and Humans,1997,27(5):553-568.
[125]Parker J R. Rank and response combination from confusion matrix data[J]. Information Fusion,2001,2(2):113-120.
[126]刘明,袁保宗,苗振江.一种双目标排序层分类器融合方法[J].自动化学报,2007,33(12)：1276-1282.
[127]Cheriet M, Kharma N, Liu C L, Suen C. Character recognition systems:a Guide for students and practitioners[M]. John Wiley and Sons,2007.
[128]Liu C L. Classifier combination based on confidence transformation[J].Pattern Recognition,2005,38(1):11-28.
[129]Dempster A P. Upper and lower probabilities induced by a multi-valued mapping[J]. Annals mathematical statistics.1967,38(2):325-339
[130]Shafer G. A mathematical theory of evidence[M]. Princeton University Press,1976.
[131]Kudo T, Matsumoto Y. Fast methods for kKernel-based text analysis[C]. Proceedings of the 41st Annual Meeting of the Association for Computational Linguistics,2003:24-31.
[2]Manning C D, Schutze H. Foundations of statistical natural language processing[M],MIT Press,1999.
[3]Daelemans W, Buchholz S, Veenstra J. Memory-based shallow parsing[C].In Proceedings of CoNLL-1999,1999:53-60.
[4]Roth D. Memory based learning in NLP[R].Technical Report 2125,Urbana, Illinois,1999.
[5]Zavrel J, Daelemans W. Memory-based learning:using similarity for smoothing[C]. Proceedings of the Thirty-Fifth Annual Meeting of the Association for Computational Linguistics and Eighth Conference of the European Chapter of the Association for Computational Linguistics,1997:436-443.
[6]Quinlan J R. C4.5:Programs for machine learning[M]. Morgan Kaufmann Publishers,1993.
[7]Heeman P A. POS tags and decision trees for language modeling[C]. SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora, College Park, Maryland, June 1999:129-137.
[8]Ratnaparkhi A. A maximum entropy model for part-of-speech tagging[C]. The 1996 Conference on Empirical Methods on Natural Language Processing,1996:133-142.
[9]Ratnaparkhi A. A liner observed time statistical parser based on maximum entropy models[C]. The 1997 Conference on Empirical Methods on Natural Language Processing, 1997:1-10.
[10]Cortes C, Vapnik V N. Support vector networks. Machine Learning[J],1995,20:273-297.
[11]Vapnik V N. Statistical learning theory[M]. Wiley-Interscience,1998.
[12]Lafferty J, McCallum A, Pereira F. Conditional random fields:Probabilistic models for segmenting and labelling sequence data[C]. Proceedings of the Eighteenth International Conference on Machine Learning,2001:282-289.
[13]Taskar B, Guestrin C, Koller D. Max-Margin markov networks[C]. Proceeding of Neural Information Processing Systems Conference,2003.
[14]Lakoff G. Hedges:a study in meaning criteria and the logic of fuzzy concepts[J]. Journal of Philosophical Logic,1973,2(4):458-508.
[15]Light M, Qiu X Y, Srinivasan P. The language of bioscience:facts, speculations, and statements in between[C]. Proceedings of the BioLINK,2004:17-24.
[16]Vincze V, Szarvas G, Farkas R, Mora G, Csirik J. The BioScope corpus:biomedical texts annotated for uncertainty, negation and their scopes[J]. BMC Bioinformatics, 2008,9(11):S9.
[17]Szarvas G. Hedge classification in biomedical texts with a weakly supervised selection of keywords[C]. Proceedings of Association for Computational Linguistics, HLT 2008: 281-289.
[18]Medlock B, Briscoe T. Weakly supervised learning for hedge classification in scientific literature [C]. The 45th Annual Meeting of the Association of Computational Linguistics, 2007:992-999.
[19]Farkas R, Vincze V, Mora G, Csirik J, Szarvas G. The CoNLL 2010 shared task:learning to detect hedges and their scope in natural language text[C]. Proceedings of the CoNLL2010 Shared Task,2010:1-12.
[20]Hyland K. Writing without conviction hedging in science research articles[J]. Applied Linguistics,1996,7:433-454.
[21]Meyers G. The pragmatics of politeness in scientific [J]. English for Specific Purposes, 2000,19(13):175-187.
[22]Meyers F. Procrustes'recipe:hedging and positivism[J]. English for Specific Purposes: An International Journal 2000,19(2):175-189.
[23]Skelton J. The representation of truth in academic medical writing[J]. Applied Linguistics,1997,18(2):121-140.
[24]Varttala, T. Remarks on the communicative functions of hedging in popular scientific and specialist research articles on medicine [J]. English for Specific Purposes,1999, 18(2):177-200.
[25]赵英玲.英语科技语体中的模糊限制语[J].外语与外语教学,1999,124(9)：15-17.
[26]冯茵,周榕.学术论文摘要中模糊限制语的调查与分析—基于英语专业毕业论文与国外期刊论文的对比研究[J].外国语言文学,2007,92(2)：108-112.
[27]Prince, E F, Frader J, Bosk C, Dipietro R J. On the hedging in physician-physician discourse[C]. In Linguistics and the Professions. J. Robert (ed.) Norwood:Ablex, 1982:83-97.
[28]喻亚男,徐畅贤.论英语模糊限制语的语用功能[J].益阳师专学报,2002,23(2)：102-103.
[29]范晓晖.医患会话中模糊限制语的语用功能[J].西北医学教育,2006,14(6)：740-742.
[30]范晓晖,孙圆圆.医学论文英文摘要模糊限制语的对比研究[J]. 西北医学教育,2008,16(6)：1187-1189.
[31]王舟.英汉学术论文摘要中模糊限制语的对比研究[J].华中科技大学学报,2006,22(6)：59-63.
[32]Kilicoglu H, Bergler S. Recognizing speculative language in biomedical research articles:a linguistically motivated perspective[C]. Proceedings of the Workshop on Current Trends in Biomedical Natural Language Processing, Columbus, Ohio, June, Association for Computational Linguistics,2008:46-53.
[33]Ganter V, Strube M. Finding hedges by chasing weasels:Hedge detection using Wikipedia tags and shallow linguistic features[C]. Proceedings of the ACL-IJCNLP 2009 Conference Short Papers, Suntec, Singapore, August, Association for Computational Linguistics, 2009:173-176.
[34]Zhou H W, Li X Y, Huang D G, Li Z Z, and Yang Y S. Exploiting multi-features to detect hedges and their scope in biomedical texts [C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:56-63.
[35]Li X X, Shen J P, Gao X, Wang X. Exploiting rich features for detecting hedges and their scope[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:36-41.
[36]Zhang S D, Zhao H, Zhou G D, Lu B L. Hedge detection and scope finding by sequence labeling with normalized feature selection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[37]Ji F, Qiu X P, Huang X J, Detecting hedge cues and their scopes with average perceptron[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July. Association for Computational Linguistics,2010:139-146.
[38]Lin Chen, Barbara Di Eugenio.A lucene and maximum entropy model based hedge detection system[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July. Association for Computational Linguistics,2010:114-119.
[39]Prabhakaran V. Uncertainty learning using SVMs and CRFs[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:132-137.
[40]Zhao Q., Sun C. J., Liu B. Q., Cheng Y.. Learning to detect hedges and their scope using CRF[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:100-105.
[41]王正群.手写体汉字识别研究[D].南京：南京理工大学,2001.
[42]Tang B Z, Wang X L, Wang X, Yuan B, Fan S X. A cascade method for detecting hedges and their scope in natural language text[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:25-29.
[43]Ozgur A, Radev D R. Detecting speculations and their scopes in scientific text[C]. Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, Singapore, August,Association for Computational Linguistics.2009:1398-1407.
[44]Velldal E,Ovrelid L, Oepen S. Resolving speculation:MaxEnt cue classification and dependency-based scope rules[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:48-55.
[45]Morante R, Asch V V, Daelemans W. Memory-based resolution of insentence scopes of hedge cues[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:48-55.
[46]Kilicoglu H, Bergler S. A high-precision approach to detecting hedges and their scopes[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[47]Zhang S D, Zhao H, Zhou G D, Lu B L. Hedge detection and scope finding by sequence labeling with procedural feature selection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:70-77.
[48]Zhu Q M, Li J H, Wang H L, Zhou G D. A unified framework for scope learning via simplified shallow semantic parsing[C]. Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing,2010:714-724.
[49]Morante R, Daelemans W. Learning the scope of hedge cues in biomedical texts[C]. Proceedings of the BioNLP 2009 Workshop, pages, Boulder, Colorado, June, Association for Computational Linguistics,2009:28-36.
[50]Rei M, Briscoe T. Combining manual rules and supervised learning for hedge cue and scope detection[C]. Proceedings of the Fourteenth Conference on Computational Natural Language Learning (CoNLL 2010):Shared Task, Uppsala, Sweden, July, Association for Computational Linguistics,2010:56-63.
[51]Joachims T. Text categorization with support vector machines:learning with many relevant features[C]. Proceedings 10th European Conference on Machine Learning(ECML), Springer Verlag,1998.
[52]Dumais S, Platt J, Heckerman D, Sahami M. Inductive learning algorithms and representations for text categorization[R]. Technical Report, Microsoft Research,1998:148-152.
[53]Lodhi H, Saunders C, Sahwe-Taylor J, Cristianini N, Watkins C. Text classification using string kernels[C]. In Neural Information Processing Systems (NIPS), MIT Press, 2001:563-569.
[54]Mitra V, Wang C J, Banerjee S. Text classification:A least square support vector machine approach[J]. Applied Soft Computing,2007,7(3):908-914.
[55]Lee K S, Kageura K. Virtual relevant documents in text categorization with support vector machines [J]. Information Processing and Management:an International Journal, 2007,43(4):902-913.
[56]Balcazar J L, Dai Yang, Watanabe O. Provably fast training algorithms for support vector machines[C]. Proceedings of IEEE International Conference on Data Mining,2001,43-50.
[57]Mangasarian O L. A finite Newton method for classification[J]. Optimization Methods and Software,2002,17(5):913-929.
[58]Keerthi S, DeCoste D. A modified finite Newton method for fast solution of large scale linear SVMs[J]. Journal of Machine Learning Research.2005,6(5):341-361.
[59]Dong J X, Krzyzak A, Suen C Y. An improved handwritten Chinese character recognition system using support vector machine[J]. Pattern Recognition Letters.2005, 26 (12):1849-1856.
[60]Dong J X, Krzyzak A, Suen C Y. Fast SVM training algorithm with decomposition on very large data sets[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2005,27(4):603-618.
[61]Wu Y C, Fan T K, Lee Y S, et al. Extracting named entities using support vector machines[J]. Lecture Notes in Bioinformatics(LNBI):Knowledge Discovery in Life Science Literature,2006,3886:91-103.
[62]Kuto T, Matsumoto Y. Chunking with support vector machines[C]. Proceeding of NAACL, 2001:192-199.
[63]Kuto T, Matsumoto Y. Japanese dependency analysis based on support vector machines[C]. In Empirical Methods in Natural Language processing and Very Large Corpora,2000:18-25.
[64]Scholkopf B. Input space versus feature dpace in kernel-based methods[J]. IEEE Translations on Neural Networks,1999,10(5):1000-1016.
[65]Muller K R. An introduction to kernel-based learning algorithms [J]. IEEE Translations on Neural Networks,2001,12(2):181-201.
[66]Boser B E, Guyon L M, and Vapnik V N. A training algorithm for optimal margin classifiers[C]. Proceedings of the 5th annual ACM workshop on computational learning theory. ACM Press,1992:144-152.
[67]Osuna E, Freund R, Girosi F. Training support vector machines:an application to face detection[C]. Proceedings CVPR'97,1997:130-136.
[68]Joachims T. Making large-scale SVM learning practical[C]. Conference on Advances in Kernel Methods Support Vector Learning, MIT Press,1999:169-184.
[69]Platt J C, Scholkopf B, Burges C, Smola A. Fast training of support vector machines using sequential minimal optimization[C]. Conference on Advances in Kernel Methods:Support Vector Learning, MIT Press,1998:185-208.
[70]Platt J C. Using analytic QP and sparseness to speed training of support vector machines[C]. Conference on Advances in neural information processing systems Ⅱ,1999:557-563.
[71]Eddy S R. Hidden markov models[J]. Current Opinion in Structural Biology,1996, 6(3):361-365.
[72]McCallum A, Freitag D, Pereira F. Maximum entropy markov models for information extraction and segmentation[C]. Proceedings of the Seventeenth International Conference on Machine Learning,2000:591-598.
[73]Kudo T, Yamamoto K, Matsumoto Y. Applying conditional random fields to Japanese morphological analysis[C]. Proceedings of EMNLP 2004,2004:230-237.
[74]黄德根,焦世斗,周惠巍.基于子词的双层CRFs中文分词[J].计算机研究与发展,2010,47(5)：962-968.
[75]周俊生,戴新宇,尹存燕,陈家骏.基于层叠条件随机场模型的中文机构名自动识别[J].电子学报,2006,34(5)：804-809.
[76]McCallum A and Li W. Early results for named entity recognition with conditional random fields, feature induction and web-enhanced lexicons[C]. Proceeding of CoNLL 2003, 2003:188-191.
[77]Sha F, Pereira F. Shallow parsing with conditional random fields[C]. Proceedings of Human Language Technology-NAACL,2003:134-141.
[78]Forney J G. The viterbi algorithm[J]. Proceedings of the IEEE,1973,61(3):268-278.
[79]Johansen S, Juselius K. Maximum likelihood estimation and inference on cointegration-with applications to the demand for money[J].Oxford Bulletin of Economics and Statistics,1990,52(2):169-210.
[80]Darroch J N, Ratcliff D. Generalized iterative scaling for log-linear models[J].The Annals of Mathematical Statistics,1972,43(5):1470-1480.
[81]Malouf R. A comparison of algorithms for maximum entropy parameter estimation[C]. Proceeding of the 6th conference on Natural language learning,2002:1-7.
[82]Shewchuk J R. An introduction to the conjugate gradient method without the agonizing pain[M]. Edition 1 1/4, School of Computer Science, Carnegie Mellon University,1994.
[83]Byrd R H, Nocedal J, Schnabel R B. Representation of quasi-Newton matrices and their use in limited memory methods[J]. Mathematical Programming,1994,63:129-156.
[84]Sha F, Pereira F. Shallow parsing with conditional random fields[C]. Proceedings of Humna Language Technology NAACL, Edmonton, Canada,2003:213-220.
[85]Collins M, Duffy N. Convolution kernels for natural language[C]. Conference on Advances in Neural Information Processing Systems,2001:625-632.
[86]Kong F, Zhou G D, Qian L H, Zhu Q M. Dependency-driven anaphoricity determination for coreference resolution[C]. Proceedings of the 23rd International Conf.on Computational Linguistics,2010:599-607.
[87]Zhou G D, Kong F. Global learning of noun phrase anaphoricity in coreference resolution via label propagation[C].Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing,2009:978-986.
[88]Zhang M, Zhang J, Su J. Exploring syntactic features for relation extraction using a convolution tree kernel[C].Proceedings of the Human Language Technology Conf.of the North American Chapter of the ACL,2006:288-295.
[89]Moschitti A, Pighin D, Basili R. Semantic role labeling via tree kernel joint inference[C]. Proceedings of the 10th Conference on Computational Natural Language Learning, New York, USA,2006:61-68.
[90]Tsuruoka Y, Tateishi Y, Kim J D, Ohta T, McNaught J, Ananiadou S, Tsujii J. Developing a robust part-of-speech tagger for biomedical text[C]. Conference on Advances in Informatics 2005:382-392.
[91]Christiane F. WordNet and wordnets[M]. Encyclopedia of Language and Linguistics, Second Edition, Oxford:Elsevier,2005:665-670.
[92]Liu Y, Tan Q, Shen K, The word segmentation rules and automatic word segmentation methods for Chinese information processing[M]. QingHua University Press and GuangXi Science and Technology Press,1994:2-36.
[93]Petrov S, Barrett L, Thibaux R, Klein D. Learning accurate, compact, and interpretable tree annotation[C]. Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics,2006:433-440.
[94]Petrov S, Klein D. Improved inference for unlexicalized parsing[C]. Proceedings of NAACL HLT 2007:404-411.
[95]Zhou G D, Zhang M, Ji D H, and Zhu Q M. Tree kernel-based relation extraction with context-sensitive structured parse tree information[C]. Proceedings of the 2007 Joint Conference on Empirical Methods in Natural Language Processing and Computational,2007:728-736.
[96]Joachims T, Cristianini N, Shawe-Taylor J. Composite kernels for hypertext categorization[C]. Proceedings of ICML-2001,2001:250-257.
[97]Zhang M, Zhang J, Su J, and Zhou G D. A composite kernel to extract relations between entities with both flat and structured features[C]. ACL 44th Proceedings of the 21st International Conference on Computational Linguistics and the 44th annual meeting of the Association for Computational Linguistics,2006:825-832.
[98]车万翔.基于核方法的语义角色标注[D].哈尔滨：哈尔滨工业大学,2008.
[99]钱龙华.命名实体间语义关系抽取研究[D].苏州：苏州大学,2009.
[100]Moschitti A. A study on convolution kernels for shallow semantic parsing[C]. Proceedings of the 42th Conference on Association for Computational Linguistic, Barcelona, Spain,2004:335-342.
[101]Zhou H W, Huang D G, Li X Y, Yang Y S. A staged and distributed strategy for bio-entity recognition[J]. International Journal of Information,14(10):3527-3536.
[102]冯志伟.自然语言处理的形式模型[M].安徽：中国科学技术大学出版社,2010.
[103]Earley J. An efficient context-free parsing algorithm[J]. Communications of the association for computing machinery,1970,13(2):94-102.
[104]宗成庆.统计自然语言处理[M].北京：清华大学出版社,2008.
[105]赵伟,戴新宇,尹存燕,陈家骏.一种规则与统计相结合的汉语分词方法[J].计算机工程与科学,2003(3)：23-25
[106]姜尚仆,陈群秀.基于规则和统计的日语分词和词性标注的研究[J].中文信息学报,2010,24(1)：117-122.
[107]周强.规则与统计结合的汉语词类标注方法[J].中文信息学报,1995,9(2)：1-10.
[108]姜柄圭,张秦龙,谌贻荣,常宝宝.面向机器辅助翻译的汉语语块自动抽取研究[J].中文信息学报,2007,21(1)：9-16.
[109]李素建,刘群,白硕.统计和规则相结合的汉语组块分析[J].计算机研究与发展,2002,39(4)：385-391.
[110]苗艳军;李军辉;周国栋.统计和规则相结合的并列结构自动识别[J].计算机应用研究,2009,26(9)：3403-3406.
[111]傅间莲,陈群秀.基于规则和统计的中文自动文摘系统[J]. 中文信息学报,2006,20(5)：10-16.
[112]昝红英,左维松,张坤丽,吴云芳.规则和统计相结合的情感分析研究[J].计算机工程与科学,2011,33(5),146-150.
[113]郭宏蕾,胡岗.统计与规则相结合的目标语言生成策略[C].2002年全国机器翻译研讨会,002：110-115.
[114]Eisele A, Federmann C, Uszkoreit H, Saint-Amand H, Kay M, Jellinghaus M, Hunsicker S, Herrmann T, Chen Y. Hybrid machine translation architectures within and beyond the EuroM atrix project[C]. Proceedings of the 12th annual conference of the European Association for Machine Translation (EAMT 2008). Hamburg, Germany,2008:27-34.
[115]徐金安.理性主义与经验主义相结合的机器翻译研究策略[J].计算机科学,2011,38(6)：223-229.
[116]Sagae K, Tsujii J. Dependency parsing and domain adaptation with LR models and parser ensembles. Proceedings of the CoNLL 2007 Shared Task[C]. Joint Conferences on Empirical Methods in Natural Language Processing and Computational Natural Language Learning (EMNLP-CoNLL'07).2007:1044-1050.
[117]Spengler A. Maximum margin markov networks for XML tag relabeling[D]. Karlsruhe: University of Karlsruhe,2005.
[118]Kleinberg E M. A Mathematically rigorous foundation for supervised learning[J]. Lecture Notes in Computer Science,2000,1857:67-76.
[119]Allein E L, Schapire R E, and Singer Y. Reducing multiclass to binary:a unifying approach for margin classifiers[J]. Journal of Machine Learning Research, 2000,1(9):113-141
[120]周俊生,戴新宇,尹存燕,陈家俊.自然语言信息抽取中的机器学习方法研究[J].计算机科学,2005,32(3)：186-189.
[121]Halteren H V, Zavrel J, Daelemans W. Improving accuracy in word class tagging through the combination of machine learning systems[J]. Coynputatioual Linguistics, 2001,27(2):199-230.
[122]Rahman A F R, Fairhurst M C. Multiple calssifier decision combination strategies for character recognition:a review[J]. International Journal on Document Analysis and Recognition.2003,5(4):166-194
[123]Raudys S, Roli F. The behavior knowledge space fusion method:analysis of generalization error and strategies for performance improvement [C]. Proceedings of 4th International Workshop on Multiple Classifier Systems (MCS). Lecture Notes in Computer Science 2709,2003:55-64.
[124]Lam L, Suen C Y. Application of majority voting to pattern recognition:an analysis of its behavior and performance[J]. IEEE Transactions on systems, man, and Cybernetics-Part A:Systems and Humans,1997,27(5):553-568.
[125]Parker J R. Rank and response combination from confusion matrix data[J]. Information Fusion,2001,2(2):113-120.
[126]刘明,袁保宗,苗振江.一种双目标排序层分类器融合方法[J].自动化学报,2007,33(12)：1276-1282.
[127]Cheriet M, Kharma N, Liu C L, Suen C. Character recognition systems:a Guide for students and practitioners[M]. John Wiley and Sons,2007.
[128]Liu C L. Classifier combination based on confidence transformation[J].Pattern Recognition,2005,38(1):11-28.
[129]Dempster A P. Upper and lower probabilities induced by a multi-valued mapping[J]. Annals mathematical statistics.1967,38(2):325-339
[130]Shafer G. A mathematical theory of evidence[M]. Princeton University Press,1976.
[131]Kudo T, Matsumoto Y. Fast methods for kKernel-based text analysis[C]. Proceedings of the 41st Annual Meeting of the Association for Computational Linguistics,2003:24-31.