基于LDA模型的Twitter中文微博热点主题词组发现

摘要

This paper proposes a method that combines LDA model and related techniques of the natural language information processing in a bid to extract hot topic phrases in Chinese microblogs in Twitter.To corroborate the effectiveness of this method,this paper selects 20,923 pieces of Chinese Tweets as samples.The extracted topic phrases are generally within expectation,demonstrating that this model is capable of recognizing hot topics and describing them.%提出了潜在狄利赫雷分布模型与自然语言信息处理相关技术相结合的一种挖掘Twitter中中文微博的热点主题词组的方法.选取了20923条中文Tweets作为样本,获取了相关热点的主题词组,与预期的效果基本吻合,表明该模型具有较好的热点识别效果和对主题进行描述的能力.