Twitter-LDA

Latent Dirichlet Allocation (LDA) has been widely used in textual analysis. The original LDA is used to find hidden "topics" in the documents, where a topic is a subject like "arts" or "education" that is discussed in the documents. The original setting in LDA, where each word has a topic label, may not work well with Twitter as tweets are short and a single tweet is more likely to talk about one topic. Hence, Twitter-LDA (T-LDA) has been proposed to address this issue. T-LDA also addresses the noisy nature of tweets, where it captures background words in tweets. As experiments in [7] have shown that T-LDA could capture more meaningful topics than LDA in Microblogs.

The original setting in Latent Dirichlet Allocation (LDA), where each word has a topic label, may not work well with Twitter as tweets are short and a single tweet is more likely to talk about one topic. Hence, Twitter-LDA (T-LDA) has been proposed to address this issue. T-LDA also addresses the noisy nature of tweets, where it captures background words in tweets.
Related Publication: Zhao, W. X., Jiang, J., Weng, J., He, J., Lim, E. P., Yan, H., & Li, X. (2011). Comparing twitter and traditional media using topic models. In Advances in Information Retrieval (pp. 338-349). http://doi.org/10.1007/978-3-642-20161-5_34

潜在狄利克雷分配（Latent Dirichlet Allocation，LDA）已被广泛应用于文本分析领域。原始LDA旨在从文档集合中挖掘隐含的“主题”，此处的主题指文档中讨论的诸如“艺术”“教育”这类核心议题。原始LDA的设定为每个词汇均对应一个主题标签，但该设定在推特场景下效果欠佳——由于单条推特篇幅极短，更倾向于仅围绕单一主题展开。为此，研究者提出了推特版潜在狄利克雷分配（Twitter-LDA，简称T-LDA）以解决该问题。T-LDA同时针对推特文本的噪声特性进行优化，能够捕捉推文中的背景词汇。已有实验[7]证实，在微博客语境下，T-LDA能够挖掘出比原始LDA更具实际意义的主题。

原始LDA的设定为每个词汇均对应一个主题标签，但该设定在推特场景下效果欠佳——由于单条推特篇幅极短，更倾向于仅围绕单一主题展开。为此，研究者提出了推特版潜在狄利克雷分配（Twitter-LDA，简称T-LDA）以解决该问题。T-LDA同时针对推特文本的噪声特性进行优化，能够捕捉推文中的背景词汇。

相关文献：Zhao, W. X., Jiang, J., Weng, J., He, J., Lim, E. P., Yan, H., & Li, X. (2011). 基于主题模型的推特与传统媒体对比研究. 载于《信息检索进展》(pp. 338-349). http://doi.org/10.1007/978-3-642-20161-5_34