Incorporating Entities in News Topic Modeling
Incorporating Entities in News Topic Modeling
复制标题
DOI：
10.1007/978-3-642-41644-6_14
复制
发表时间：
2013-11
期刊：
影响因子：
--
通讯作者：
Linmei Hu;Juan-Zi Li;Zhihui Li;Chao Shao;Zhixing Li
中科院分区：
文献类型：
--
作者：
Linmei Hu;Juan-Zi Li;Zhihui Li;Chao Shao;Zhixing Li
新闻文章通过关注新闻中的人物、时间和地点等命名实体来表达信息。然而，从大量的新闻文章中提取实体、单词和主题之间的关系是非常重要的。像Latent Dirichlet Allocation这样的主题建模方法在文本分析中被广泛应用于挖掘隐藏主题，并取得了相当好的效果。但是，它不能显式地显示单词和实体之间的关系。本文提出了一种生成模型——以实体为中心的主题模型（entity - centered Topic model， ECTM），将实体主题作为词主题的混合体来总结实体、词和主题之间的相关性。在真实新闻数据集上的实验表明，我们的模型比最先进的实体主题模型（corlda2）具有更低的困惑度和更好的实体聚类。我们还对ECTM的结果进行了分析，并进一步与CorrLDA2进行了比较。
News articles express information by concentrating on named entities like who, when, and where in news. Whereas, extracting the relationships among entities, words and topics through a large amount of news articles is nontrivial. Topic modeling like Latent Dirichlet Allocation has been applied a lot to mine hidden topics in text analysis, which have achieved considerable performance. However, it cannot explicitly show relationship between words and entities. In this paper, we propose a generative model, Entity-Centered Topic Model(ECTM) to summarize the correlation among entities, words and topics by taking entity topic as a mixture of word topics. Experiments on real news data sets show our model of a lower perplexity and better in clustering of entities than state-of-the-art entity topic model(CorrLDA2). We also present analysis for results of ECTM and further compare it with CorrLDA2.