基于最大熵模型的汉语标点句缺失话题自动识别初探

J4 ›› 2015, Vol. 37 ›› Issue (12): 2282-2293.

基于最大熵模型的汉语标点句缺失话题自动识别初探

卢达威1，宋柔2

(1.北京大学中国语言文学系，北京 100871；2.北京语言大学语言信息处理研究所，北京 100083)

收稿日期:2015-09-01 修回日期:2015-11-05 出版日期:2015-12-25 发布日期:2015-12-25
基金资助:
国家自然科学基金资助项目(61171129);国家973计划资助项目(2014CB340502)

Automatic recognition of the absent topics in Chinese
punctuation clauses based on maximum entropy model

LU Dawei1,SONG Rou2

(1.Department of Chinese Language and Literature,Peking University，Beijing 100871;
2.Institute of Language Information Processing,Beijing Language and Culture University，Beijing 100083,China)

Received:2015-09-01 Revised:2015-11-05 Online:2015-12-25 Published:2015-12-25

摘要/Abstract

摘要：

本文的任务是判别标点句缺失话题是上句的主语还是宾语，将该任务作为标点句缺失话题自动识别研究的切入点。首先归纳了判别这一任务的一系列字面特征和语义特征，然后结合规则和最大熵模型，进行自动判别实验。结果显示，对特定类别动词的实验F值达到82%。对实验结果的分析说明，动词特征和语义特征对判别该任务的作用最大，规则方法和统计方法在判别任务中不能偏废，精细化的知识对判别的性能有重要影响。

关键词: 广义话题结构, 新支话题, 自动识别, 最大熵模型

Abstract:

We focus on the task of the automatic recognition,which identify whether an absent topic of a punctuation clause is the subject or object of its previous sentence. We regard this task as the pointcut of the automatic recognition of absent topics in Chinese punctuation clauses. Several literal features and semantic features are summerized to achieve this task by combining the rules and the maximum entropy model. Experimental results show that Fscore of this recognition approach reaches 82% for the samples of some specific verbs. Experimental results analysis shows that verb features and semantic features play the most important role in the recognition process; neither rules nor statistics can be neglected, and refined knowledge has great influence on the performance of the recognition .

Key words: generalized topic structure;new branch topic;automatic recognition;maximum entropy model

卢达威1，宋柔2. 基于最大熵模型的汉语标点句缺失话题自动识别初探[J]. J4, 2015, 37(12): 2282-2293.

LU Dawei1,SONG Rou2. Automatic recognition of the absent topics in Chinese
punctuation clauses based on maximum entropy model [J]. J4, 2015, 37(12): 2282-2293.

参考文献

［1］Song Rou.The delesion of the fronts of clauses in Chinese narratives ［J］.Journal of Chinese Information Processing,1992,6(3):6268.(in Chinese)
［2］Song Rou.Research on properties of syntactic relation between Pclauses in modern Chinese ［J］.Chinese Teaching in the World,2008(2):2644.(in Chinese)
［3］Song Rou.Stream model of generalized topic structure in Chinese text［J］.Studies of The Chinese Language,2013(6):483494.(in Chinese)
［4］Lu Dawei,Song Rou,Shang Ying.Cognitive complexity of topic in Chinese text based on generalized topic structure theory ［J］.Journal of Chinese Information Processing,2014,28(5):112124.(in Chinese)
［5］Lu Dawei.The cognitive and computational research of generalized topic structure in Chinese discourse:A corpusdriven approach ［D］.Beijing:Beijing Language and Culture University,2015.(in Chinese)
［6］Jiang Yuru,Song Rou.Topic clause identification based on generalized topic theory ［J］.Journal of Chinese Information Processing,2012,26(5):114119.(in Chinese)
［7］Jiang Yuru,Song Rou.Topic clause identification method based on specific features ［J］.Journal of Computer Applications,2014,36(5):13451349.(in Chinese)
［8］Jiang Yuru,Song Rou.Optimization of candidate topic clause evaluation function in topic clause indentification ［J］.Jouranl of Beijing University of Technology,2014,40(1):4348.(in Chinese)
［9］Shang Ying.Empirical research of generalized topic structure theory ［D］.Beijing:Beijing Language and Culture University,2014.(in Chinese)
［10］Ji Cui,Lu Dawei,Song Rou.Research on pragmatic function of verbs addressing new branch topic ［J］.Journal of Chinese Information Processing,2014,28(3):2227.(in Chinese)
［11］Zhang Ruipeng.The research about the constraint rules of syntax relation in crosspuctuation sentence in written mordern Chinese ［M］.Beijing:China Social Sciences Press,2013.(in Chinese)
［12］Zhang Bin. Descriptive Grammar of Modern Chinese［M］.Beijing:The Commerical Press, 2010.(in Chinese)
［13］Berger A L, Pietra V J D, Pietra S AD. A maximum entropy approach to natural language processing ［J］.Computational Linguistics，1996,22(1):3971.
附中文参考文献：
［1］宋柔.汉语叙述文中的小句前部省略现象初析［J］.中文信息学报,1992,6(3):6268.
［2］宋柔.现代汉语跨标点句句法关系的性质研究［J］.世界汉语教学,2008(2):2644.
［3］宋柔.汉语篇章广义话题结构的流水模型［J］.中国语文,2013(6):483494.
［4］卢达威,宋柔,尚英.从广义话题结构考察汉语篇章话题的认知复杂度［J］.中文信息学报,2014,28(5):112124.
［5］卢达威.语料库实证的汉语篇章广义话题结构认知和计算研究［D］.北京：北京语言大学,2015.
［6］蒋玉茹,宋柔.基于广义话题理论的话题句识别［J］.中文信息学报,2012,26(5):114119.
［7］蒋玉茹,宋柔.基于细粒度特征的话题句识别方法［J］.计算机应用,2014,34(5):13451349.
［8］蒋玉茹,宋柔.话题句识别中候选话题句评估函数的优化［J］.北京工业大学学报,2014,40(1):4348.
［9］尚英.汉语篇章广义话题结构理论的实证性研究［D］.北京：北京语言大学,2014.
［10］季翠,卢达威,宋柔.动词引出新支话题的语用功能研究［J］.中文信息学报,2014,8(3):2227.
［11］张瑞朋.现代汉语书面语中跨标点句句法关系约束条件的研究［M］.北京：中国社会科学出版社,2013.
［12］张斌.现代汉语描写语法［M］.北京:商务印书馆,2010.

[1]	夏吾吉1,2，华却才让1. 基于投射的藏语语义依存分析研究[J]. 计算机工程与科学, 2019, 41(10): 1868-1873.
[2]	卢达威1,宋柔2,尚英3. 汉语篇章广义话题结构的认知计算模型研究[J]. 计算机工程与科学, 2018, 40(07): 1264-1274.
[3]	杨美妮1，何涛2，沈静1，张建军1. 不规则文本中商品名称识别的特征选择[J]. 计算机工程与科学, 2016, 38(10): 2153-2157.
[4]	丁兆颖1，姚迪2，吴琳2，毕经平2，赵瑞莲1. 一种基于改进的DBSCAN的面向海量船舶位置数据码头挖掘算法[J]. J4, 2015, 37(11): 2061-2067.
[5]	张志民，欧建平，皇甫堪. 模拟通信信号调制方式自动识别算法[J]. J4, 2013, 35(3): 163-167.
[6]	才藏太. 基于最大熵分类器的藏文句子边界自动识别方法研究[J]. J4, 2012, 34(6): 187-190.
[7]	张晓艳王挺陈火旺. 基于混合统计模型的汉语命名实体识别方法[J]. J4, 2006, 28(6): 135-139.
[8]	张茹[1] 刘明业[2]. 二维条码在信息安全领域的应用研究[J]. J4, 2004, 26(2): 108-109.
[9]	窦勤颖姚青. 条码技术的发展及其应用[J]. J4, 2003, 25(5): 50-52.