基于Bi-LSTM的古籍事件句触发词分类方法研究

doi:10.13998/j.cnki.issn1002-1248.21-0262

摘要/Abstract

摘要： [目的/意义]开展面向数字人文的古籍触发动词识别及分类研究,对于古籍文本的深层次挖掘和内容揭示具有重大的意义。本文利用深度学习分类算法,探索依据古籍触发词进行事件句文本多元分类的自动化方法。[方法/过程]在构建了典籍事件触发词分类体系和触发词典的基础上,选取4个不同类别的事件句文本作为实验数据,利用Onehot和Tokenizer对类别标签和句子文本进行分别编码后,输入Bi-LSTM模型中训练分类器,并通过调整参数设置了对比实验,采取通用的评价指标分析了分类器的性能。[结果/结论]经过多次训练和调整之后得到的分类器,在测试集的评估中精确度达到了0.95,证明基于深度学习的实验方法和构建的触发词数据集能够有效的帮助我们实现古籍事件句文本的自动化多元分类。

关键词: 触发词分类, Bi-LSTM模型, 多元分类, 《左传》

Abstract: [Purpose/Significance] It is of great significance to carry out research on the recognition and classification of trigger verbs in ancient books oriented to digital humanities for the deep mining and content revealing of ancient texts. This paper uses the deep learning classification algorithm to explore an automated method for multivariate classification of event sentence text based on trigger words in ancient books. [Method/Process] Based on the construction of the classic event trigger word classification system and trigger dictionary, four different types of event sentence texts are selected as experimental data, and the category labels and sentence texts are coded separately using Onehot and Tokenizer, and then the classifier is trained in the Bi-LSTM model, and a comparative experiment is set by adjusting the parameters, and the performance of the classifier is analyzed by using a general evaluation index. [Results/Conclusions] The classifier after many training and adjustments has an accuracy of 0.95 in the evaluation of the test set, which proves that the experimental method based on deep learning and the constructed trigger word data set can effectively help us realize automatic multivariate classification of event sentence text of ancient books.

Key words: trigger word classification, Bi-LSTM model, multivariate classification, Zuo Zhuan

中图分类号:

G350

马晓雯, 何琳, 刘建斌, 李章超, 高丹. 基于Bi-LSTM的古籍事件句触发词分类方法研究[J]. 农业图书情报学报, 2021, 33(9): 27-36.

MA Xiaowen, HE Lin, LIU Jianbin, LI Zhangchao, GAO Dan. The Trigger Verb Classification Method of Event Sentences in Ancient Chinese Classics Based on Bi-LSTM[J]. Journal of Library and Information Science in Agriculture, 2021, 33(9): 27-36.

参考文献

[1] 魏晓萍. 数字人文背景下数字化古籍的深度开发利用[J]. 农业图书情报学刊, 2018, 30(9): 106-110.
WEI X P.In-depth development and utilization of digital ancient books under the background of digital humanities[J]. Journal of agricultural library and information science, 2018, 30(9): 106-110.
[2] 邓三鸿, 胡昊天, 王昊, 等. 古文自动处理研究现状与新时代发展趋势展望[J]. 科技情报研究, 2021, 3(1): 1-20.
DENG S H, HU H T, WANG H, et al.Research status of automatic processing of ancient texts and prospects for development trends in the new era[J]. Science and technology information research, 2021, 3(1): 1-20.
[3] 苏晓丹. 开放域事件抽取关键技术研究[D]. 郑州: 解放军信息工程大学, 2017.
SU X D.Research on key technologies of open domain event extraction[D]. Zhengzhou: PLA information engineering university, 2017.
[4] 张建海. 基于深度学习的生物医学事件抽取研究[D]. 大连: 大连理工大学, 2016.
ZHANG J H.Research on biomedical event extraction based on deep learning[D]. Dalian: Dalian university of technology, 2016.
[5] 张仲华, 苏方方, 姬东鸿. 生物医学事件触发词识别研究[J]. 计算机应用研究, 2017, 34(3): 661-664, 670.
ZHANG Z H, SU F F, JI D H.Research on trigger word recognition of biomedical events[J]. Application research of computers, 2017, 34(3): 661-664, 670.
[6] 何馨宇. 基于文本挖掘的生物事件抽取关键问题研究[D]. 大连:大连理工大学, 2019.
HE X Y.Research on key issues of biological event extraction based on text mining[D]. Dalian: Dalian university of technology, 2019.
[7] 贾遂民, 张玉, 张腾飞. 一种基于介词用法的灾难事件信息抽取方法[J]. 计算机与现代化, 2015(7): 116-119.
JIA S M, ZHANG Y, ZHANG T F.A disaster event information extraction method based on the usage of prepositions[J]. Computer and modernization, 2015(7): 116-119.
[8] 邱奇志, 周三三, 刘长发, 等. 基于文体和词表的突发事件信息抽取研究[J]. 中文信息学报, 2018, 32(9): 56-65, 74.
QIU Q Z, ZHOU S S, LIU C F, et al.Research on the extraction of emergency information based on style and vocabulary[J]. Chinese journal of information, 2018, 32(9): 56-65, 74.
[9] 王宁, 陈湧, 郭玮, 等. 基于知识元的突发事件案例信息抽取方法[J]. 系统工程, 2014, 32(12): 133-139.
WANG N, CHEN Y, GUO W, et al.Information extraction method of emergencies based on knowledge element[J]. Systems engineering, 2014, 32(12): 133-139.
[10] 蒋德良. 基于规则匹配的突发事件结果信息抽取研究[J]. 计算机工程与设计, 2010, 31(14): 3294-3297.
JIANG D L.Research on information extraction of emergencies based on rule matching[J]. Computer engineering and design, 2010, 31(14): 3294-3297.
[11] 万齐智, 万常选, 胡蓉, 等. 基于句法语义依存分析的中文金融事件抽取[J]. 计算机学报, 2021, 44(3): 508-530.
WAN Q Z, WAN C X, HU R, et al.Chinese financial event extraction based on syntactic and semantic dependency analysis[J]. Chinese journal of computers, 2021, 44(3): 508-530.
[12] 罗明, 黄海量. 基于词汇-语义模式的金融事件信息抽取方法[J]. 计算机应用, 2018, 38(1): 84-90.
LUO M, HUANG H L.Financial event information extraction method based on vocabulary-semantic model[J]. Computer applications, 2018, 38(1): 84-90.
[13] 李江龙, 吕学强, 周建设, 等. 金融领域的事件句抽取[J]. 计算机应用研究, 2017, 34(10): 2915-2918, 2945.
LI J L, LU X Q, ZHOU J S, et al.Event sentence extraction in the financial field[J]. Application research of computers, 2017, 34(10): 2915-2918, 2945.
[14] 赵小明, 朱洪波, 陈黎, 等. 基于多分类器的金融领域多元关系信息抽取算法[J]. 计算机工程与设计, 2011, 32(7): 2348-2351.
ZHAO X M, ZHU H B, CHEN L, et al.Multi-classifier-based multi-relation information extraction algorithm in the financial field[J]. Computer engineering and design, 2011, 32(7): 2348-2351.
[15] 薛聪, 高能, 查达仁, 等. 事件库构建技术综述[J]. 信息安全学报, 2019, 4(2): 83-106.
XUE C, GAO N, CHA D R, et al.Overview of event library construction technology[J]. Journal of information security, 2019, 4(2): 83-106.
[16] 高强, 游宏梁. 事件抽取技术研究综述[J]. 情报理论与实践, 2013, 36(4): 114-117, 128.
GAO Q, YOU H L.A review of event extraction technology re-search[J]. Information theory and practice, 2013, 36(4):114-117, 128.
[17] 轩小星, 廖涛, 高贝贝. 中文事件触发词的自动抽取研究[J]. 计算机与数字工程, 2015, 43(3): 457-461.
XUAN X X, LIAO T, GAO B B.Research on automatic extraction of Chinese event trigger words[J]. Computer and digital engineering, 2015, 43(3): 457-461.
[18] 谭红叶. 中文事件抽取关键技术研究[D]. 哈尔滨: 哈尔滨工业大学, 2008.
TAN H Y.Research on key technologies of Chinese event extraction[D]. Harbin: Harbin institute of technology, 2008.
[19] 李培峰, 周国栋, 朱巧明. 基于语义的中文事件触发词抽取联合模型[J]. 软件学报, 2016, 27(2): 280-294.
LI P F, ZHOU G D, ZHU Q M.A joint model of Chinese event trig-ger word extraction based on semantics[J]. Journal of software, 2016, 27(2): 280-294.
[20] 项威, 王邦. 中文事件抽取研究综述[J]. 计算机技术与发展, 2020, 30(2): 1-6.
XIANG W, WANG B.A review of Chinese event extraction research[J]. Computer technology and development, 2020, 30(2): 1-6.
[21] 吴敏. 网络短文本主题聚类研究[D]. 武汉: 华中科技大学, 2015.
WU M.Research on topic clustering of internet short texts[D].Wuhan: Huazhong university of science and technology, 2015.
[22] 徐鑫鑫, 刘彦隆, 宋明. 利用加权词句向量的文本相似度计算方法[J]. 小型微型计算机系统, 2019, 40(10): 2072-2076.
XU X X, LIU Y L, SONG M.Text similarity calculation method using weighted words and sentence vectors[J]. Small microcomputer system, 2019, 40(10): 2072-2076.
[23] 李岩. 基于深度循环神经网络的关系抽取方法研究[D]. 开封: 河南大学, 2019.
LI Y.Research on relation extraction method based on deep recurrent neural network[D]. Kaifeng: Henan university, 2019.
[24] 梁杰, 陈嘉豪, 张雪芹, 等. 基于独热编码和卷积神经网络的异常检测[J]. 清华大学学报(自然科学版), 2019, 59(7): 523-529.
LIANG J, CHEN J H, ZHANG X Q, et al.Anomaly detection based on one-hot encoding and convolutional neural network[J]. Journal of Tsinghua university (natural science edition), 2019, 59(7): 523-529.
[25] 易士翔, 尹宏鹏, 郑恒毅. 基于BiLSTM的公共安全事件触发词识别[J]. 工程科学学报, 2019, 41(9): 1201-1207.
YI S X, YIN H P, ZHENG H Y.Trigger word recognition of public safety events based on BiLSTM[J]. Journal of engineering science, 2019, 41(9): 1201-1207.
[26] 王立荣. Word2vec-CNN-Bi-LSTM短文本情感分类[J]. 福建电脑, 2020, 36(1): 11-16.
WANG L R.Word2vec-CNN-Bi-LSTM short text sentiment classification[J]. Fujian computer, 2020, 36(1): 11-16.
[27] 和志强, 杨建, 罗长玲. 基于Bi-LSTM神经网络的特征融合短文本分类算法[J]. 智能计算机与应用, 2019, 9(2): 21-27.
HE Z Q, YANG J, LUO C L.Feature fusion short text classification algorithm based on Bi-LSTM neural network[J]. Intelligent computers and applications, 2019, 9(2): 21-27.
[28] 阳萍, 谢志鹏. 基于Bi-LSTM模型的定义抽取方法[J]. 计算机工程, 2020, 46(3): 40-45.
YANG P, XIE Z P.Definition extraction method based on Bi-LSTM model[J]. Computer engineering, 2020, 46(3): 40-45.
[29] CHEN Y, XU L, LIU K.Event extraction via dynamicmulti-pooling convolutional neural networks[C]//Proceedings of the 53rd ACL and the 7th IJCNLP. Beijing, China: ACL, 2015: 167-176.
[30] NGUYEN T H, CHO K, GRISHMAN R.Joint event extraction via recurrent neural networks[C]//Proceedings of the 2016 NAACL: Human language technologies. San Diego, California: ACL, 2016: 300-309.