摘要
根据对网络舆情分析的需求,构建出基于聚类的网络舆情热点发现及分析系统。通过对样本网页文本的特征提取,构建向量空间模型,使用OPTICS算法获取网页热点簇,根据热点簇特征向量对网页进行二次聚类,从而获取关于舆情的时间演变模式,为相关领域研究提供决策支持。通过二次聚类,提高舆情网页相关度的质量,使网络舆情分析更为准确可靠。
According to the requirement of online public opinion analysis, this paper builds an online public opinion hotspot detection and analysis system based on document clustering. It builds vector space model by abstracting document features from sample Web pages, and get the hot - spot cluster by OPTICS algorithm. According the vector of hot - spot cluster, the Web pages are clustered for the second time. At last, it gets the time evolution mode about the public opinion to afford decision support for specific field, and improves the quality of page correlation and analyze the public opinion more accurately.
出处
《现代图书情报技术》
CSSCI
北大核心
2009年第3期74-79,共6页
New Technology of Library and Information Service
基金
教育部人文社会科学研究项目"互联网舆情信息分析与管理机制研究"(项目编号:08JC870003)的研究成果之一
关键词
网络舆情
热点发现
舆情分析
文本聚类
Online public opinion Hotspot detection Public opinion analysis Document clustering
