基于词向量和条件随机场的煤矿安全事故本体概念抽取  

Concept Extraction of Coal Mine Safety Accident Ontology Based on Word Embeddings and CRFs

作　　者：刘思含 刘旭红[1,2] 刘秀磊[1,3] 李红臣[4] LIU Si-han;LIU Xu-hong;LIU Xiu-lei;LI Hong-chen(Computer School,Beijing Information Science and Technology University,Beijing 100101,China;Beijing Key Laboratory of Internet Cuhure and Digital Dissemination Research,Beijing Information Science and Technology University,Beijing 100192,China;Laboratory of Data Science and Competitive Intelligence,Beijing Information Science and Technology University,Beijing 100101,China;Communication Information Center,State Administration of Work Safety,Beijing 100013,China)

机构地区：[1]北京信息科技大学计算机学院,北京100101 [2]北京信息科技大学网络文化与数字传播北京市重点实验室,北京100192 [3]北京信息科技大学数据科学与情报分析实验室,北京100101 [4]国家安全生产监督管理总局通信信息中心,北京100013

出　　处：《煤炭技术》2018年第9期178-181,共4页Coal Technology

基　　金：国家重点研发计划资助(2016YFC0801407);国家自然科学基金资助(61601039);北京信息科技大学科研基金资助(1625008);网络文化与数字传播北京市重点实验室开放课题资助(ICDDXN006);网络与交换技术国家重点实验室(北京邮电大学)开放课题资助(SKLNST-2016-2-08);北京信息科技大学软件工程专业学位点建设项目资助(5121723402);北京市教育委员会科技计划面上项目(KM201811232018)

摘　　要：煤矿安全事故的"人-机-环-管"信息种类繁多、结构复杂。基于传统条件随机场模型的概念抽取方法只考虑了词语的统计特征,因此效果不够理想。为解决上述问题,提出一种基于词向量和条件随机场模型的概念抽取方法。该方法改进了词向量模型并综合考虑了统计特征和语义相似性特征。通过对比实验,证明了其具有更好的性能。＂People-Machine-Environment-Management＂ information of coal mine safety accident has various kinds and complex structure. The concept extraction method based on the traditional conditional random field model considers only the statistical characteristics of words. So its effect is not ideal. In order to solve the above problem, a concept extraction method based on word vector and conditional random field model is proposed. The method improved word embedding model and considers both statistical features and semantic similarity features. It is proved that it has better performance by comparative experiment.

关 键 词：煤矿安全事故本体 条件随机场 词向量 语义相似度 概念抽取 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]