面向历史灾害地震的Web信息精确抽取与分析方法

A Method of Accurate Extraction and Analysis of Web Data on Historical Disaster Earthquakes

摘要:

以中国大陆地区灾害地震目录为基础，选取2010—2019年灾害地震的互联网信息，提出基于百度搜索引擎的信息获取技术，并以“时间、地名、震级”为关键词，设计一套URL生成规则。使用该技术进行百度检索，得到前100个站点的主体文字信息，建立地震信息基础语料库，形成灾害地震的网络灾情信息获取方法；通过采用已有的停用词词库剔除无用信息，对爬取到的信息进行初步清洗工作，进一步深入挖掘隐含信息，探索灾害关联关系，为震后互联网灾情信息快速获取建立基础。

Abstract:

Taking the earthquake catalogs from the Internet information of earthquakes between 2010 to 2019 in mainland China as an example, we propose an information acquisition technique based on Baidu search engine, and generate a set of URL generation rules with "time, place name, magnitude" as keywords. The first 100 sites retrieved by Baidu by using this technique are used to build a basic corpus of earthquake information and to form a method for acquiring Internet disaster information on earthquakes. The existing deactivation thesaurus is used to eliminate useless information, and then to conduct preliminary cleaning of the crawled information. The further digging into the implied information is performed in order to explore disaster correlations, and to establish a basis for rapid acquisition of Internet disaster information after earthquakes.