基于典型相关分析的多视图跨领域情感分类.pdfVIP

第34 卷 第24 期 计 算 机 工 程 2010 年12 月 Vol.34 No.24 Computer Engineering December 2010 ·人工智能及识别技术· 文章编号：1000—3428(2010)24—0186—03 文献标识码：A 中图分类号：TP391 基于典型相关分析的多视图跨领域情感分类 黄贤立 (淮阴师范学院计算机科学与技术学院，江苏 淮安 223300) 摘 要：跨领域的文本分类，是指利用有标记领域的知识去帮助另一个概率分布不同的，未标记领域的知识进行分类的问题。从多视图学 习的视角提出一个新的跨领域文本分类的方法(MTV 算法) 。通过在核空间典型相关分析中引入与标记相关的信息，MTV 算法可以得到一 个判别性能更优的公共子空间。在多个情感类文本数据上的实验表明，MTV 算法可以大大提升传统监督式学习算法面对领域迁移时的分 类性能，并且在引入判别式的核空间典型相关分析后，进一步优化性能。 关键词：跨领域学习；迁移学习；情感分类；文本分类；核空间典型相关分析 Multi-view Cross Domain Sentiment Classification Based on Canonical Correlation Analysis HUANG Xian-li (School of Computer Science and Technology, Huaiyin Normal University, Huaian 223300, China) 【Abstract 】Cross domain text classification aims at leveraging the knowledge in labeled source domains to predict the unlabeled data in another domain, where the distribution is different from the source domains. This paper proposes a new algorithm for cross domain classification in a multi-view learning perspective, named MulTi-View(MTV) transfer classification. By incorporating discriminative information related to labels, a more discriminative common low dimensional space is extracted. In MTV perspective, MTV algorithm provides a clear and neat way to transfer knowledge between domains. By applying KCCA, MTV also gives a closed form solution. Experiments show that MTV algorithm can significantly improve the performance of traditional supervised learn