摘要
针对国内事件新闻语料处理问题,提出了一种基于地名字典与朴素贝叶斯方法的事件新闻发生地点抽取方法。该方法分为两个阶段,利用地名字典初步筛选,通过机器学习提取新闻发生地点的表述特征,从而实现地名抽取。算法结合地名之间的行政所属关系,引入匹配因子,提高精确度。实验结果表明,该方法的精确率和召回率分别为95.12%和90.19%,且易于实现,对其他新闻文本信息挖掘具有一定的借鉴意义。
Aiming at the problem of event news corpus processing in China, a method of event news location extraction based on place name dictionary and naive Bayesian method is proposed. This method can be divided into two stages, using the preliminary screening of the place names dictionary, and extracting the expression features of news occurrence sites through machine learning, thus realizing the place names extraction. The algorithm combines the administrative affiliation between place names and introduces matching factors to improve accuracy. The experimental results show that the accuracy and recall rates of this method are 95.12% and 90.19% respectively, and it is easy to implement, which has a certain reference value for other news text information mining.
作者
刘佳琪
罗永莲
Liu Jiaqi;Luo Yonglian(School of Computer Engineering and Science, Shanghai University, Shanghai 200444, China;School of Information Technology & Engineering, Jinzhong University, Jinzhong Shanxi 030619, China)
出处
《信息与电脑》
2019年第15期53-54,57,共3页
Information & Computer
基金
山西省教育科学“十三五”规划课题（项目编号：GH-18091）
关键词
地名抽取
地名字典
朴素贝叶斯模型
地名规则
事件新闻
toponym extraction
toponym dictionary
naive bayesian model
toponym rules
emergency news
