基于联合非负矩阵分解的话题变迁检测方法

doi:10.3969/j.issn.1000-3428.2018.01.006

计算机工程

基于联合非负矩阵分解的话题变迁检测方法

陈梦伟¹,吕钊¹,崔修涛²

(1.华东师范大学计算机科学与技术系,上海 200241; 2.上海长江计算机有限公司,上海 200241)

收稿日期:2016-12-21 出版日期:2018-01-15 发布日期:2017-01-15
作者简介:陈梦伟(1987—)，男，硕士研究生，主研方向为大数据分析、知识处理；吕钊，副教授；崔修涛，高级工程师、博士。
基金资助:
上海市科学技术委员会科研计划项目(16511102702);上海市经济和信息化委员会项目（150643）。

Topic Change Detection Method Based on Joint Nonnegative Matrix Factorization

CHEN Mengwei ¹,LU Zhao ¹,CUI Xiutao ²

(1.Department of Computer Science and Technology, East China Normal University, Shanghai 200241, China; 2.Shanghai Changjiang Computer Co.,Ltd.,Shanghai 200241,China)

Received:2016-12-21 Online:2018-01-15 Published:2017-01-15

摘要/Abstract

摘要： 在大规模时序文档集中,异同话题缺乏从时序文档集中识别跟踪分析话题随时间变迁的能力。为此,提出一种面向时序文档语料库的话题变迁检测方法。该方法从时序文档语料库中发现相似话题和异同话题。利用改进的联合非负矩阵分解算法,从多个数据集中提取话题集合。为避免引入噪声话题,计算所有话题的话题熵,以获取优质话题,并通过运用词云和趋势图来分析话题变迁趋势。在20Newsgroups和LTN2011数据集上的实验结果表明,该方法可以有效地从时序文档集中发现异同话题,且提取的话题效果好、准确率高。

关键词: 联合非负矩阵分解, 话题模型, 时序异同话题, 优质话题, 话题变迁检测

Abstract: In large-scale temporal documents similarities and differences do not have the ability to identily topics from temporal documents and to track and analyze topics over time.To this end,a method of topic change detection for temporal document corpus is proposed.Similar topics and similarities and foundations are found in the temporal document corpus.Using the improved joint Nonnegative Matrix Factorization(NMF) algorithm,similarities and differences were found in the the timeseries document.To avoid the introduction of noise topics,by calculating the topic of all topic entropy,in order to obtain high-quality topics.Use the word cloud and trend graph to analyze the trend of topic change.Experimental results of two real data sets,20Newsgroups and LTN2011 show that this method can effectively find similarities and differences from the tempord of documents,and the extraction topic is effect and the accuracy is high.

Key words: Joint Nonnegative Matrix Factorization (NMF), topic model, temporal similarities and differences topic, high quality topic, topic change detection

中图分类号:

TP391

陈梦伟,吕钊,崔修涛. 基于联合非负矩阵分解的话题变迁检测方法[J]. 计算机工程, doi: 10.3969/j.issn.1000-3428.2018.01.006.

CHEN Mengwei,LU Zhao,CUI Xiutao. Topic Change Detection Method Based on Joint Nonnegative Matrix Factorization[J]. Computer Engineering, doi: 10.3969/j.issn.1000-3428.2018.01.006.

https://www.ecice06.com/CN/Y2018/V44/I1/35

参考文献

参考文献［1］楚克明,李芳.基于LDA模型的新闻话题的演化［J］.计算机应用与软件,2011,28(4):4-7. ［2］ESTER M,KRIEGEL H P,SANDER J,et al.A Density-based Algorithm for Discovering Clusters in Large Spatial Databases with Noise［C］//Proceedings of KDD’96.Berlin,Germany:Springer,1996:226-231. ［3］JIANG Y,LI X,MENG W.Discword:Learning Discriminative Topics［C］//Proceedings of IEEE/WIC/ACM International Joint Conferences on Web Intelligence and Intelligent Agent Technologies.Washington D.C.,USA:IEEE Press,2014:63-70. ［4］ZHANG H,KIM G,XING E P.Dynamic Topic Modeling for Monitoring Market Competition from Online Text and Image Data［C］//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York,USA:ACM Press,2015:1425-1434. ［5］LIU Y,NICULESCU M A,GRYC W.Topic-link Lda:Joint Models of Topic and Author Community［C］//Proceedings of the 26th Annual International Conference on Machine Learning.New York,USA:ACM Press,2009:665-672. ［6］KALYANAM J,MANNTRACH A,SAEZ T D,et al.Leveraging Social Context for Modeling Topic Evolution［C］//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York,USA:ACM Press,2015:517-526. ［7］王鑫,李璐,王晓芳.基于 Nystrm 谱聚类的词典学习［J］.计算机工程与应用,2014,50(6):112-117. ［8］LEE D D,SEUNG H S.Algorithms for Non-negative Matrix Factorization［C］//Proceedings of Advances in Neural Information Processing Systems.［S.1.］:MIT Press,2001:556-562. ［9］KIM H,CHOO J,KIM J,et al.Simultaneous Discovery of Common and Discriminative Topics via Joint Nonnegative Matrix Factorization［C］//Proceedings of the 21th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.New York,USA:ACM Press,2015:567-576. ［10］LI Q,CHEN B,MA S,et al.Contrastive Ttopic Evolution Discovery via Nonnegative Matrix Factorization［C］//Proceedings of IEEE International Conference on Communications.Washington D.C.,USA:IEEE Press,2016:1-6. ［11］BADEA L.Extracting Gene Expression Profiles Common to Colon and Pancreatic Adenocarcinoma Uing Simultaneous Nonnegative Matrix Factorization［J］.Biocomputing,2008,290(13):279-290. ［12］KIM J,MONTEIRO R,PARK H.Group Sparsity in Nonnegative Matrix Factorization［C］//Proceedings of SDM’12.［S.1.］:SIAM Press,2012:851-862. ［13］KIM J,HE Y,PARK H.Algorithms for Nonnegative Matrix and Tensor Factorizations:A Unified View Based on Block Coordinate Descent Framework［J］.Journal of Global Optimization,2014,58(2):285-319. ［14］李谦.非负矩阵分解及其在高维数据应用中的研究［D］.北京:北京交通大学,2014. ［15］ESTEVEZ P A,TESMER M,PEREZ C A,et al.Normalized Mutual Information Feature Selection［J］.IEEE Transactions on Neural Networks,2009,20(2):189-201. ［16］BLEI D M,NG A Y,JORDAN M I.Latent Dirichlet Allocation［J］.Journal of Machine Learning Research,2003,3(1):993-1022. ［17］LIU H,WU Z,CAI D,et al.Constrained Nonnegative Matrix Factorization for Image Representation［J］.IEEE Transactions on Pattern Analysis and Machine Intelligence,2012,34(7):1299-1311. ［18］WANG Q,CAO Z,XU J.Group Matrix Factorization for Scalable Topic Modeling［C］//Proceedings of the 35th International ACM SIGIR Conference on Research and Development in Information Retrieval.New York,USA:ACM Press,2012:375-384. ［19］舒振球,赵春霞.基于局部学习的受限非负矩阵分解算法［J］.华中科技大学学报(自然科学版),2015,43(7):82-86. ［20］杜世强,石玉清,王维兰,等.基于图正则化的半监督非负矩阵分解［J］.计算机工程与应用,2012,48(36):194-200. ［21］蓝龙.半监督非负矩阵分解算法及其在文本聚类中的应用［D］.长沙:国防科学技术大学,2012. 编辑索书志

[1]	余冲,李晶,孙旭东,傅向华. 基于词嵌入与概率主题模型的社会媒体话题识别[J]. 计算机工程, 2017, 43(12): 184-191.
[2]	杨春明，张晖，石大文. BBS网络舆情的在线自适应话题演化模型[J]. 计算机工程, 2014, 40(7): 317-321.

选择文件类型/文献管理软件名称

选择包含的内容

基于联合非负矩阵分解的话题变迁检测方法

Topic Change Detection Method Based on Joint Nonnegative Matrix Factorization

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价

模态框（Modal）标题

选择文件类型/文献管理软件名称

选择包含的内容

基于联合非负矩阵分解的话题变迁检测方法

Topic Change Detection Method Based on Joint Nonnegative Matrix Factorization

PDF

可视化

被引次数

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 2

编辑推荐

Metrics

本文评价