基于Web文本的灾害事件信息获取进展

大数据时代海量网络文本中蕴含的灾害事件信息是防灾减灾研究和应用的重要资源。从异构的Web文本中快速、准确抽取灾害事件时空信息和属性信息,分析其时空动态变化模式与趋势并进行可视化表达,是当前地理信息与灾害信息领域关注的热点。本文从Web文本挖掘的整体技术框架、灾害主题页面抓取、灾害事件信息解析与抽取、灾害事件信息空间展示分析、以及防灾减灾应用系统等方面调研和综述了相关进展。针对防灾减灾领域的Web文本信息获取未来趋势,分析概括了全流程的Web文本灾害事件信息提取适用技术,并指出未来研究趋势：① 重点开展全球灾害信息全景式获取分析,实现全球灾害事件信息的自动化获取、分析及可视化展示;② 向联合国可持续发展目标（SDGs）和中国“一带一路”倡议,加强典型热点区域的Web灾害事件信息获取分析应用研究并形成示范系统;③ 按照数据、信息、知识的应用层次,建立以大数据挖掘和分析技术支撑的新型防灾减灾知识服务系统。

Progress in Information Acquisition of Disaster Events from Web Texts

In this era of big data transfer and use, the extraction of disaster event information from huge quantities of network data is important to facilitate research on disaster prevention and reduction. In comparison with traditional disaster information, disaster information based on Web text is dynamic, heterogeneous, and massive, has space-time aspects, and accesses multiple sources. How to extract and visualize the spatio temporal and attribute information of disaster events from Web text, and track dynamic change patterns and trends of such events over space and time, is a growing area of research in geographic and disaster information systems. This study reviews the progress of relevant researchs including network data mining technology frameworks, disaster theme web page crawling, the extraction of disaster event information, the visualization and spatial distribution characteristics analysis of disaster events and the application system for disaster prevention and reduction. By examining the trend of disaster information acquisition for disaster prevention and reduction from the internet, this study analyzed and summarized the appropriate technologies of information extraction from Web text and discussed the development trends in the following three aspects: (1) Focusing on global disaster information acquisition and analysis. The fundamental trend is to realize the automatic acquisition, analysis, and visualization of global disaster event information to ensure disaster prevention and reduction. (2) To realize the United Nations' 2030 Agenda for Sustainable Development and China's "the Belt and Road" strategy, strengthening of the disaster event information analysis research and its application to typical regions is one of the research hotspots in the field of Web disaster information acquisition and application. (3) Establishing a new disaster prevention and reduction knowledge service system supported by big data mining and analysis technologies according to the application level of data, information, and knowledge will be one of the future research trends.