计算机工程 ›› 2019, Vol. 45 ›› Issue (1): 153-158.doi: 10.19678/j.issn.1000-3428.0049801

• 人工智能及识别技术 • 上一篇    下一篇

基于卷积双向长短期记忆网络的事件触发词抽取

陈斌1,周勇1,刘兵1,2   

  1. 1.中国矿业大学 计算机科学与技术学院,江苏 徐州 221116; 2.中国科学院电子研究所,北京 100094
  • 收稿日期:2017-12-21 出版日期:2019-01-15 发布日期:2019-01-15
  • 作者简介:陈斌(1993—),男,硕士研究生,主研方向为机器学习、数据挖掘;周勇(通信作者),教授、博士生导师;刘兵,副教授
  • 基金项目:

    国家自然科学基金青年基金“面向高维数据的稀疏非参核学习方法研究”(61403394);国家自然科学基金面上项目“多目标低秩非参核学习模型与优化方法研究”(61572505)

Event Trigger Word Extraction Based on Convolutional Bidirectional Long Short Term Memory Network

CHEN Bin 1,ZHOU Yong 1,LIU Bing 1,2   

  1. 1.College of Computer Science and Technology,China University of Mining and Technology,Xuzhou,Jiangsu 221116,China; 2.Institute of Electrics,Chinese Academy of Sciences,Beijing 100094,China
  • Received:2017-12-21 Online:2019-01-15 Published:2019-01-15

摘要:

传统事件触发词抽取方法在特征提取过程中过分依赖自然语言处理工具,容易造成误差累积。为解决该问题,在卷积双向长短期记忆网络的基础上,提出一种事件触发词抽取方法。通过卷积操作提取单词上下文语境信息,同时利用长短期记忆网络保留句子级别特征,从而提高事件触发词的抽取性能。在ACE2005英文语料上的实验结果表明,该方法在事件触发词识别与分类阶段的F值达到69.5%,具有较好的抽取性能

关键词: 事件抽取, 触发词, 卷积神经网络, 循环神经网络, 自然语言处理, 特征提取

Abstract:

The traditional event trigger word extraction methods over rely on Natural Language Processing(NLP) tools in the feature extraction process,which is easy to cause error accumulation.To solve this problem,a new event trigger extraction based on convolutional bidirectional Long Short Term Memory(LSTM)network is proposed.This method utilizes a convolution operation to extract word contextual information and retains sentence-level features through LSTM unit to improve the performance of event trigger word extraction.Experimental results on the ACE2005 English corpus show that the method achieves an F value of 69.5% on the event trigger word recognition and classification stage,which has a good extraction effect.

Key words: event extraction, trigger word, Convolutional Neural Network(CNN), Recurrent Neural Network(RNN), Natural Language Processing(NLP), feature extraction

中图分类号: