融合标准知识的事故灾难领域词典构建  

Construction of Disaster Domain Dictionary Integrating Standard Knowledge

作　　者：伊然 张甜 邢心羽 马雯雯 张鲲洋 刘文玲 YI Ran;ZHANG Tian;XING Xin-yu;MA Wen-wen;ZHANG Kun-yang;LIU Wen-ling(School of Management,China University of Mining and Technology(Beijing))

机构地区：[1]中国矿业大学(北京)管理学院

出　　处：《中国标准化》2022年第15期88-94,117,共8页China Standardization

基　　金：国家重点研发计划项目“典型灾害情景下突发事件应对数字化关键标准研究与应用”(项目编号:2021YFF0600400);“中国矿业大学(北京)大学生创新训练项目”资助。

摘　　要：现阶段我国正处于事故灾难的频发时期,目前国内针对这一领域的专业化词典相对欠缺。本文以有关事故灾难领域的国家应急标准及网络数据为语料库,首先对语料库进行预处理,其次利用TF-IDF算法筛选出种子词集合,同时借助深度学习中的Word2Vec模型进行词向量训练,然后利用相似度计算的方式确定出领域候选词,最后融合应急标准术语完成事故灾难领域词典的构建。经SVM分类器与朴素贝叶斯分类器验证,将本文所构建的领域词典加入到结巴自定义词典后,分类器在准确率、召回率与F1值上分别提高了11%、11%、12%以及5%、8%、6%,表明本文构建的领域词典质量较好。China now is in a period of frequent disasters,but there is a lack of the professional dictionary for this important field.This paper takes the national emergency standards and network data in the field of disasters as the corpus,and on the basis of corpus preprocessing,it selects Seed word through the TF-IDF algorithm,trains the word vector with the help of the Word2 Vec model in deep learning,and then determines the domain candidate by calculating the similarity of the vector,and finally integrates the emergency standard terms to construct the disaster domain dictionary.Verified by SVM classifier and naive Bayes classifier,after adding the domain dictionary constructed in this paper to the stutter userdefined dictionary,the rate s of accuracy,recall,and F1 value of the classifier are improved by 11%,11%,12% and 5%,8%,6% respectively,indicating that the domain dictionary has a high quality.

关 键 词：事故灾难 TF-IDF算法 Word2Vec模型 SVM分类器 朴素贝叶斯分类器 

分 类 号：X4[环境科学与工程—灾害防治] TP391.1[自动化与计算机技术—计算机应用技术]