基于LDA思想,用先离散时间型话题模型思路建立网络突发事件话题的空间模型,并根据特征词之间的关联度使用社会网络分析方法进行话题演化路径的分析。实验表明,此思路能够较全面体现话题演化路径,为网络突发事件分析提供有效途径。
Based on LDA model, we use the time pre-discretized method to analyze the topic space, then according to the relation degree of feature words we analyze the migration of topic content by social net- work analysis method. The proposed method was experimentally verified to be efficient for detecting topic evolution of internet emergency.