Chinese Named Entity Recognition Based on Bidirectional LSTM
发布时间：2021-12-30 22:04
命名实体识别是文本信息处理中一项关键且基础的工作,其任务是识别出待处理文本中具有实际意义的实体。它在自然语言处理中是一项基础性的工作,在信息抽取,机器翻译,自动问答等领域有着广泛的应用。评价一个命名实体是否被正确识别包括两点:实体的边界是否正确;实体的类型是否标注正确。本文在时间和精力有限的情况下主要是识别出数据库中的人名、地名、组织名这三类命名实体,而且这三类也是最难而且最重要的三类命名实体。命名实体识别的准确度对后续相关的工作有深刻的影响。目前已经有研究这方面内容的专门机构,Conll-2002就有专门的命名实体识别专题。目前中文的命名实体识别与英文的相比,中文的命名实体识别更加的困难,主要是因为英语的命名实体具有一些明显的区分标志,比如英文中的首字母要大写以及英文单词之间要空格,所以它的实体边界比较容易确定,剩下的任务只需要去确定实体的类别。和英文相比,中文命名实体识别任务更加复杂,中文实体边界的确定更加的困难。汉语命名实体识别的难点主要存在于:中文文本中没有像英文那样很明显的空格之类的边界标示符;中文中存在外国人名译名和地名译名这些特殊实体类型；中文中不可能用一个统一的模型来描...
【文章来源】：华中师范大学湖北省 211工程院校 教育部直属院校
【文章页数】：59 页
【学位级别】：硕士
【文章目录】：
Acknowledgements
abstract
1 Introduction
1.1 Research Background and Significance of the Subject
1.1.1 Research Background
1.1.2 Research Significance
1.2 Analysis of Research Status at Home and Abroad
1.2.1 Status of Foreign Research
1.2.2 Status of Chinese Research
1.3 Problems and Analysis of Named Entity Recognition
1.4 Main Work of This Thesis
1.5 Structure of Thesis
2 Related Work
2.1 Method of Named Entity Recognition
2.1.1 Hidden Markov Model
2.1.2 Conditional Random Field Model
2.2 Basic Neural Network
2.2.1 Neurons
2.2.2 Perceptron and Multilayer Network
2.2.3 Recurrent Neural Network
2.3 Tensorflow Framework
2.4 Summary of This Chapter
3 BiLSTM-CRF Models for Chinese Named Entity Recognition
3.1 LSTM Networks
3.2 Bidirectional LSTM Networks
3.3 Conditional Random Field Networks
3.4 BiLSTM-CRF Networks
3.4.1 Chinese Data Preprocessing
3.4.2 Word Vector Layer
3.4.3 BiLSTM Layer
3.4.4 CRF layer
4 Experiments
4.1 Experimental Environment
4.2 Data Sets
4.2.1 MSRA Corpus
4.2.2 People's Daily Corpus
4.3 Evaluation Methods
4.4 Experimental Results
5 Conclusion and Future Work
5.1 Conclusion
5.2 Future Work
References
Appendix A 摘要
【参考文献】：
期刊论文
[1]Research of Clinical Named Entity Recognition Based on Bi-LSTM-CRF[J]. 秦颖,曾颖菲. Journal of Shanghai Jiaotong University（Science）. 2018(03)
[2]基于BI-LSTM-CRF模型的中文分词法[J]. 张子睿,刘云清. 长春理工大学学报(自然科学版). 2017(04)
[3]命名实体识别研究[J]. 张晓艳,王挺,陈火旺. 计算机科学. 2005(04)
[4]基于层叠隐马模型的汉语词法分析[J]. 刘群,张华平,俞鸿魁,程学旗. 计算机研究与发展. 2004(08)
本文编号：3559046
【文章来源】：华中师范大学湖北省 211工程院校 教育部直属院校
【文章页数】：59 页
【学位级别】：硕士
【文章目录】：
Acknowledgements
abstract
1 Introduction
1.1 Research Background and Significance of the Subject
1.1.1 Research Background
1.1.2 Research Significance
1.2 Analysis of Research Status at Home and Abroad
1.2.1 Status of Foreign Research
1.2.2 Status of Chinese Research
1.3 Problems and Analysis of Named Entity Recognition
1.4 Main Work of This Thesis
1.5 Structure of Thesis
2 Related Work
2.1 Method of Named Entity Recognition
2.1.1 Hidden Markov Model
2.1.2 Conditional Random Field Model
2.2 Basic Neural Network
2.2.1 Neurons
2.2.2 Perceptron and Multilayer Network
2.2.3 Recurrent Neural Network
2.3 Tensorflow Framework
2.4 Summary of This Chapter
3 BiLSTM-CRF Models for Chinese Named Entity Recognition
3.1 LSTM Networks
3.2 Bidirectional LSTM Networks
3.3 Conditional Random Field Networks
3.4 BiLSTM-CRF Networks
3.4.1 Chinese Data Preprocessing
3.4.2 Word Vector Layer
3.4.3 BiLSTM Layer
3.4.4 CRF layer
4 Experiments
4.1 Experimental Environment
4.2 Data Sets
4.2.1 MSRA Corpus
4.2.2 People's Daily Corpus
4.3 Evaluation Methods
4.4 Experimental Results
5 Conclusion and Future Work
5.1 Conclusion
5.2 Future Work
References
Appendix A 摘要
【参考文献】：
期刊论文
[1]Research of Clinical Named Entity Recognition Based on Bi-LSTM-CRF[J]. 秦颖,曾颖菲. Journal of Shanghai Jiaotong University（Science）. 2018(03)
[2]基于BI-LSTM-CRF模型的中文分词法[J]. 张子睿,刘云清. 长春理工大学学报(自然科学版). 2017(04)
[3]命名实体识别研究[J]. 张晓艳,王挺,陈火旺. 计算机科学. 2005(04)
[4]基于层叠隐马模型的汉语词法分析[J]. 刘群,张华平,俞鸿魁,程学旗. 计算机研究与发展. 2004(08)
本文编号：3559046
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/3559046.html
