一种解决“中心主题湮没问题”的基于图模型的Labeled-LDA文本分类算法

计算机科学 ›› 2014, Vol. 41 ›› Issue (3): 223-227.

一种解决“中心主题湮没问题”的基于图模型的Labeled-LDA文本分类算法

李伟,马永征,沈一

中国科学院计算机网络信息中心北京100190;中国科学院计算机网络信息中心北京100190;中国科学院计算机网络信息中心北京100190

出版日期:2018-11-14 发布日期:2018-11-14
基金资助:
本文受中科院十二五信息化项目“科研信息化应用推进工程”(XXH12503)资助

Labeled-LDA Text Classification Algorithm Based on Graph Model for “Central Topic Oblivion Problem”

LI Wei,MA Yong-zheng and SHEN Yi

Online:2018-11-14 Published:2018-11-14

摘要/Abstract

摘要： 隐含狄利克雷分配(LDA,Latent Dirichlet Allocation)是一种用于挖掘文档集中潜在主题信息的无监督主题模型。而LDA模型的变形Labeled-LDA则可以作为有监督的多标签分类器,它建立了主题与标签的一一映射,从而学习出词与标签之间的关系。近年来,图模型在文本挖掘方面的应用取得了良好的效果,通过对文档建立图模型,为进一步分析文档的语义提供了新的途径。提出了一种利用Labeled-LDA和文档图模型进行文本分类的新算法,与传统的LDA模型方法相比,该方法的性能有较大的提高。

关键词: 文本分类,图挖掘,图模型,隐含狄利克雷分配中图法分类号TP391.1文献标识码A

Abstract: Latent Dirichlet Allocation(LDA) is an unsupervised topic model used to mining potential topic information from the corpus．Labeled-LDA as a mutation of LDA can be used to do multi-classification on labeled documents,which establishes the one-to-one mapping from topic to label and learns the relationship between words and labels．Recently,the application of graph model has obtained good results in text mining,which provides a new way to analyze semantics of documents．This paper proposed a new method combining complex network theory and Labeled-LDA to do text classification．The experimental results show that our new method gets an improvement according to Macro_F1compared to the traditional LDA model.

Key words: Text classification,Graph mining,Graph model,LDA

李伟,马永征,沈一. 一种解决“中心主题湮没问题”的基于图模型的Labeled-LDA文本分类算法[J]. 计算机科学, 2014, 41(3): 223-227. https://doi.org/

LI Wei,MA Yong-zheng and SHEN Yi. Labeled-LDA Text Classification Algorithm Based on Graph Model for “Central Topic Oblivion Problem”[J]. Computer Science, 2014, 41(3): 223-227. https://doi.org/

参考文献

[1] 苏金树,张博锋,徐昕．基于机器学习的文本分类技术研究进展[J]．软件学报,2006,17:1848-1859
[2] Chen L,Tokuda N,Nagai A．A new differential LSI space-based probabilistic document classifier[J]．Information Processing Letters,2003,88(5):203-212
[3] Hofmann T．Probabilistic Latent Semantic Indexing[C]∥SI-GIR．1999:50-57
[4] 李文波,孙乐,张大鲲．基于Labeled-LDA模型文本分类新算法[J]．计算机学报,2008,31:620-627
[5] Blei D,Ng A,Jordan M．Latent Dirichlet Allocation[J]．Journal of Machine Learning Research,2003,3:993-1022
[6] Ramage D,Hall D,Nallapati R,et al．Labeled LDA:A supervised topic model for credit attribution in multi-labeled corpora[C]∥Proceedings of the 2009Conference on Empirical Methods in Natural Language Processing．August 2009:248-256
[7] 黄云平,孙乐,李文波．基于上下文图模型文本表示的文本分类研究[C]∥第四届全国信息检索与内容安全学术会议论文集(上)．2008
[8] 赵鑫,李晓明．主题模型在文本挖掘中的应用[R]．PKU-CS-NCIS-TR2011XX．June 2011
[9] Griffiths T L,Steyvers M．Finding scientific topics[C]∥Proceedings of the National Academy of Sciences．April 2004,1:5228-5235
[10] Griffiths T．Gibbs sampling in the generative model of Latent Dirichlet Allocation．http://people.cs.umass.edu/～wallach/courses/s11/cmpsci791ss/readings/griffithso2gibbs.pdf
[11] Chang C-C,Lin C-J．LIBSVM:a library for support vector machines．http://www.csie.ntu.edu.tw/~cjlin/libsvm,2001
[12] Blei D M．Probabilistic topic models[J]．Communications of the ACM,2012,5:77-84
[13] Blei D M,McAuliffe J D．Supervised topic models[C]∥NIPS．2007
[14] Cancho R F I,Sole R V．The small world of human language[J]．Proceedings of The Royal Society of London B:Biological Sciences,2001,8(1482):2261-2265
[15] Salton G,Buckley C．Term-weighting approaches in automatic text retrieval[J]．Information Processing & Management,1988,4(5):513-523

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed