基于微格式的分布式网络舆情监测系统  

Distributed Network Public Sentiment Monitoring System Based on Micro-formats

作　　者：赵永升[1] 

机构地区：[1]鲁东大学信息科学与工程学院,山东烟台264025

出　　处：《计算机工程》2013年第11期272-275,共4页Computer Engineering

基　　金：山东省自然科学基金资助项目(ZR2011FM017);山东省教育厅科研基金资助项目(J08LJ01)

摘　　要：为解决集中式舆情监测系统中的误报率较高、响应不及时等问题,提出一种基于微格式的分布式舆情监测系统。给出该系统的组织结构模型、数据的组织结构和处理流程。根据话题的功率谱估计实现基于微格式的网络监测热点发现,构建基于成分的聚类预测模型对当前热点话题进行管理,并选择合适的应对措施。仿真实验结果证明,与SmartC系统相比,该系统具有较高的舆情处理精度和响应处理效率。Slow response rate and high false alarm rate are two problems in public sentiment monitoring system centers. In order to deal with them, a novel system is presented based on the micro-formats. And its distributed organization model, data structure and work flows are given as following. Power spectrum estimation detection algorithms and micro-formats selection methods are utilized to monitor public sentiment hotspots. And a novel predication model can get their trends to select beforehand cases. Simulation experimental results show that the system has higher public sentiment precision and efficiency compared with SmartC system.

关 键 词：微格式 分布式 舆情 话题检测 功率谱估计 聚类 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]