基于微博的灾害信息快速提取方法研究

Disaster Information Extraction from Microblog

提出了一个灾害微博的实时处理框架,框架基于Spark流计算平台,利用从灾害历史微博数据中提取的信息对获取到的灾害微博进行实时分类统计。以2017年九寨沟地震和林芝地震为例,基于爬虫、文本预处理、文本分类模型等方法实现了灾害微博实时处理原型系统,并通过实验验证了原型系统的可用性。

This paper introduces a real-time processing framework for disaster related microblog. This framework conducts real-time classification and aggregation on disaster related microblog data using information extracted from history microblog data.It consists of three modules: data collecting module, history data processing module and real-time data processing module, which is based on a stream computing platform.To validate this framework,this paper develops a prototype system using several technologies such as web crawling, text pre-processing and text classification models. Finally, this paper takes 2017 Jiuzhaigou earthquake and Linzhi earthquake as cases to test this prototype system.