Multi-view learning via probabilistic latent semantic analysis

详细信息	查看全文 | 推荐本文 |

• 作者：Fuzhen Zhuang^a ; ^b ; ^{zhuangfz@ics.ict.ac.cn} ; George Karypis^c ; ^{karypis@cs.umn.edu} ; Xia Ning^c ; Qing He^a ; ^{heq@ics.ict.ac.cn} ; Zhongzhi Shi^a ; ^{shizz@ics.ict.ac.cn}
• 关键词：Multi-view learning ; Generative model ; Probabilistic Latent Semantic Analysis (PLSA)
• 刊名：Information Sciences
• 出版年：2012
• 期刊代码：62_00200255
• 类别：cp
• 出版时间：15 September, 2012
• 卷：199
• 期：Complete
• 页码：20-30
• 文件大小：314 K

摘要

Multi-view learning arouses vast amount of interest in the past decades with numerous real-world applications in web page analysis, bioinformatics, image processing and so on. Unlike the most previous works following the idea of co-training, in this paper we propose a new generative model for Multi-view Learning via Probabilistic Latent Semantic Analysis, called MVPLSA. In this model, we jointly model the co-occurrences of features and documents from different views. Specifically, in the model there are two latent variables y for the latent topic and z for the document cluster, and three visible variables d for the document, f for the feature, and v for the view label. The conditional probability p(z鈭?em>d), which is independent of v, is used as the bridge to share knowledge among multiple views. Also, we have p(y鈭?em>z, v) and p(f鈭?em>y, v), which are dependent of v, to capture the specifical structures inside each view. Experiments are conducted on four real-world data sets to demonstrate the effectiveness and superiority of our model.