多视图有监督的LDA模型

doi:10.3969/j.issn.0372-2112.2014.10.026

电子学报 ›› 2014, Vol. 42 ›› Issue (10): 2040-2044.DOI: 10.3969/j.issn.0372-2112.2014.10.026

多视图有监督的LDA模型

李晓旭¹, 李睿凡^2,3, 冯方向², 曹洁¹, 王小捷^2,3

1. 兰州理工大学计算机与通信学院, 甘肃兰州 730050;
2. 北京邮电大学计算机学院, 北京 100876;
3. 教育部信息网络工程研究中心, 北京 100876

收稿日期:2013-05-02 修回日期:2013-10-05 出版日期:2014-10-25 发布日期:2014-10-25
作者简介:李晓旭 女,1982年生于吉林白城.兰州理工大学计算机与通信学院讲师.研究方向为计算机视觉、机器学习. E-mail:xiaoxulilut@gmail.com;李睿凡 男,1975年生于河北完县.北京邮电大学计算机学院讲师.主要研究兴趣为多模态智能信息处理. E-mail:rfli@bupt.edu.cn
基金资助:
国家自然科学基金(No.61263031);甘肃省自然科学基金(No.1310RJZA034);中央高校基本科研业务费专项资金(No.2013RC0304)

Multi-view Supervised Latent Dirichlet Allocation

LI Xiao-xu¹, LI Rui-fan^2,3, FENG Fang-xiang², CAO Jie¹, WANG Xiao-jie^2,3

1. College of Computer and Communication, Lanzhou University of Technology, Lanzhou, Gansu 730050, China;
2. School of Computer Science, Beijing University of Posts and Telecommunications, Beijing 100876, China;
3. Engineering Research Center of Information Networks, Ministry of Education.Beijing 100876, China

Received:2013-05-02 Revised:2013-10-05 Online:2014-10-25 Published:2014-10-25

摘要/Abstract

摘要：

本文主要关注多视图数据的分类问题.考虑到集成分类方法可组合多个弱分类器构成一个强分类器,以及主题模型能学习复杂数据的语义表示,本文试图将集成学习思想引入主题模型中,以便同时学习多视图数据的分类规则和预测性语义特征.具体地,结合概率主题模型LDA模型和集成分类方法Softmax混合模型,提出了一个多视图有监督的分类模型.基于变分EM方法,推导了该模型的参数估计算法.两个真实图像数据集上的实验结果表明了提出模型有较好的分类性能.

关键词: 多视图分类, 概率主题模型, 变分期望最大化

Abstract:

In the paper, we mainly focus on classifition on multi-view data.Considering that ensemble methods can combine weak classifiers to construct a strong classifier, and topic model can learn latent representations from complex data, we try to introduce ensemble idea to topic model, such that predictive latent representation could be obtained and multi-view classifier could be learned.We propose multi-view supervised latent Dirichlet allocation (multi-view sLDA) model by combining latent Dirichlet allocation model and the mixture of softmax model which is an ensemble classification model.Moreover, we derive a parameter estimation algorithm of the proposed model based on variational expectation maximization (EM) procedure.The experimental results on two real datasets show the effectiveness of the proposed model.

Key words: multi-view classification, probabilistic topic model, variational expectation maximization

中图分类号:

TP181

李晓旭, 李睿凡, 冯方向, 曹洁, 王小捷. 多视图有监督的LDA模型[J]. 电子学报, 2014, 42(10): 2040-2044.

LI Xiao-xu, LI Rui-fan, FENG Fang-xiang, CAO Jie, WANG Xiao-jie. Multi-view Supervised Latent Dirichlet Allocation[J]. Acta Electronica Sinica, 2014, 42(10): 2040-2044.

参考文献

[1] Wu L,et al.Multimodal integration—A statistical view[J].IEEE Transactions on Multimedia,1999,1(4):334-341.
[2] Wang G,et al.Building text features for object image classification[A].IEEE Conference on Computer Vision and Pattern Recognition (CVPR-09)[C].IEEE,2009.1367-1374.
[3] Blei D M and Jordan M I.Modeling annotated data[A].Proc of the 26th Annual International ACM SIGIR Conference on Research and Development in Informaion Retrieval[C].ACM,2003.127-134.
[4] Frawley W J,et al.Knowledge discovery in databases:An overview[J].AI Magazine,1992,13(3):57.
[5] Memisevic,R.On multi-view feature learning[A].Proceedings of the 29th International Conference on Machine Learning (ICML-12)[C].New York,USA:Omnipress,2012.161-168.
[6] Chen N,et al.Predictive subspace learning for multi-view data:A large margin approach[A].Advances in Neural Information Processing Systems[C].Vancouver:Curran Associates,2010.361-369.
[7] Wang C,et al.Simultaneous image classification and annotation[A].IEEE Conference on Computer Vision and Pattern Recognition (CVPR-09)[C].IEEE,2009.1903-1910.
[8] Sonnenburg S,et al.Large scale multiple kernel learning[J].The Journal of Machine Learning Research,2006,7:1531-1565.
[9] Blei D M,et al.Latent dirichlet let allocation[J].The Journal of Machine Learning Research,2003,3:993-1022.
[10] Bishop C M,et al.Pattern Recognition and Machine Learning[M].Springer New York,2006.461-674.
[11] Braun M,McAuliffe J.Variational inference for large-scale models of discrete choice[J].Journal of the American Statistical Association,2010,105(489):324-335.
[12] Russell B C,et al.Labelme:A database and web-based tool for image annotation[J].International Journal of Computer Vision,2008,77(1):157-173.
[13] Li L J,Fei-Fei L.What,where and who? classifying events by scene and object recognition[A].IEEE 11th International Conference on Computer Vision(ICCV 2007)[C].IEEE,2007.1-8.
[14] Cawley G C,et al.Sparse multinomial logistic regression via bayesian L1 regularisation[A].Advances in neural information processing systems[C].Cambridge MA USA:MIT Press,2007.209-216.
[15] Chang CC,Lin C J.Libsvm:A library for support vector machines[J].ACM Transactions on Intelligent Systems and Technology (TIST),2011,2(3):27.

多视图有监督的LDA模型

Multi-view Supervised Latent Dirichlet Allocation

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	王雪松, 张翰林, 程玉虎. 基于自编码器和超图的半监督宽度学习系统[J]. 电子学报, 2022, 50(3): 533-539.
[2]	高俊涛, 王梅, 徐光会, 刘聪. 正则语言推断综述[J]. 电子学报, 2021, 49(12): 2479-2489.
[3]	孙新, 盖晨, 申长虹, 张颖捷. 基于短语向量和主题加权的关键词抽取方法[J]. 电子学报, 2021, 49(9): 1682-1690.
[4]	孟海宁, 冯锴, 朱磊, 张贝贝, 童新宇, 黑新宏. 基于Laplacian图谱的短文本聚类算法[J]. 电子学报, 2021, 49(9): 1716-1723.
[5]	周东明, 张灿龙, 李志欣, 王智文. 基于多层级视觉融合的图像描述模型[J]. 电子学报, 2021, 49(7): 1286-1290.
[6]	唐焕玲, 郑涵, 刘艳红, 马思源, 窦全胜, 鲁明羽. Tr-SLDA:一种面向交叉领域的迁移主题模型[J]. 电子学报, 2021, 49(3): 605-613.
[7]	金苍宏, 董腾然, 陈天翼, 吴明晖, 李国军, 周胜利. 融合序列分解与时空卷积的时序预测算法[J]. 电子学报, 2021, 49(2): 233-238.
[8]	汪成亮, 郑诚, 曾卓. 基于软件定义智能的睡眠动作识别[J]. 电子学报, 2021, 49(1): 85-89.
[9]	王格格, 郭涛, 余游, 苏菡. 基于生成对抗网络的无监督域适应分类模型[J]. 电子学报, 2020, 48(6): 1190-1197.
[10]	王宇, 陈晨, 张家琦, 彭丝雨, 高军峰. 任务提示的脑认知加工过程中神经振荡机制研究[J]. 电子学报, 2020, 48(5): 960-965.
[11]	王金艳, 胡春, 牛当当, 李先贤. 基于IAPS的扩展规则局部搜索算法[J]. 电子学报, 2020, 48(5): 899-905.
[12]	牛当当, 吕帅, 王金艳, 刘斌. 基于MACR和CAL启发式的求差知识编译算法[J]. 电子学报, 2020, 48(2): 285-290.
[13]	方佳艳, 刘峤. 具有同步化特征选择的迭代紧凑非平行支持向量聚类算法[J]. 电子学报, 2020, 48(1): 44-58.
[14]	牛当当, 吕帅, 王金艳. 基于MOVR启发式的求差知识编译算法[J]. 电子学报, 2019, 47(11): 2299-2303.
[15]	余游, 冯林, 王格格, 徐其凤. 一种基于伪标签的半监督少样本学习模型[J]. 电子学报, 2019, 47(11): 2284-2291.