基于 Web 新闻的案（事）件抽取与时空分析系统

摘要

News narration has been an important way to understand the social dynamics since ancient times ,in a big data era , because of the objectivity and authenticity of web news itself , it contains a large number of data value .For the rich, easy collection and other advantage of the Crime event information in Web news , this study developed an event extraction and Spatio -temporal analysis system based on the Web information .Capture the crime event information of Fuzhou on web news sites .Distinguish and analysis of the Web news, the method of support vector machine was used to classify the crime event text .The accuracy of multi classification was 75%. At last, do spatiotemporal extraction and spatiotemporal analysis of crime event .Take the drug crime event as an example;the kernel density method was used to predict the crime concentration from the extraction result and data from security department .The result shows that the overall gathering areas of drug crimes are located in the Chayuan police district and Xiangyuan police district .It is good to the public security department and social dynamic analysis in Fuzhou .%新闻，自古以来便是人们了解社会动态的重要途径，大数据时代，由于Web新闻自身所具有的客观性和真实性，其蕴含的数据价值凸显。针对新闻网站中案（事）件信息丰富、易采集等优点，研究开发一套基于Web新闻的案（事）件抽取与时空分析系统，抓取各个新闻网站对发生于福州的案（事）件相关信息的报道，对新闻信息进行判别清洗与解析，采用支持向量机进行案（事）件类别分类，多类别分类精度达75％，抽取经分类处理之后的案（事）件文本中的案（事）件时空信息并进行时空分析，以毒品案（事）件为例，将解析结果与公安毒品案（事）件分别做核密度估计，结果表明，福州毒品事件集中发生于茶园派出所和象园派出所等辖区。该系统有利于分析福州社会动态，也为公安部门提供了信息辅助。