一种基于时序主题模型的网络热点话题演化分析系统

Study on a Hot Topics Analysis System based on Time Sliced Topic Model

设计基于网络数据的时序主题演化系统Hot Topics Analysis System(HTAS),实现网络热点话题数据源自动定制、自动获取和自动存储。针对中文话题分析,HTAS集成了Google的开源分词系统IKAnalyzer,批量处理中文文档。采用LDA模型对网络热点话题主题进行提取,并利用时间标签发现热点话题,通过图标形式可视化展示其演化规律。通过"钓鱼岛事件"的实验证明,HTAS系统能快速获取、存储网络热点话题,并能有效地分析其演化趋势。

A Hot Topics Analysis System （HTAS） based on time sliced network data was proposed. HTAS realized the network hot topic data source automatically collected, acquisition and storage. HTAS integrated the google revenue segmentation system IKAnalyzer to batch processing of Chinese documents. HTAS used LDA model to extract and time label to find the evolution of the hot topics on the network. Experiments of Diaoyudao as the hot event show that, the system can effectively acquire, store and analyze this hot topic evolution trend.