Automatic Text Categorization Based on Transfer Knowledge Method

摘要：

This paper introduces a new strategy for acquiring transfer knowledge and how to apply it to text categorization tasks. We focus on how to acquire transfer knowledge from the already labeled data and use this knowledge to help the classification tasks in other feature spaces or domains. By extracting linguistic information such as part-of-speech and cooccurrence of keywords, we form the representation of our transfer knowledge base. Through experiments on cross-corpus classification tasks and crossvalidation over single corpus, and comparing the results of them, we demonstrate the effectiveness of our method.

关键词： Transfer Learning Transfer Knowledge Text Categorization Co-occurrence Heterogeneous Feature Space

作者: Geli Fei Dequan Zheng

作者单位: MOE-MS Key Laboratory of Natural Language Processing and Speech Harbin Institute of Technology Harbin, China

会议类型: 国际会议

会议名称: 2010 Third Pacific-Asia Conference on Web Mining and Web-based Application(2010年第三届web挖掘和基于web应用亚太会议 WMWA 2010)

会议地点: 桂林

会议语种:英文

页码: 1-4

在线出版日期: 2010-11-17（万方平台首次上网日期，不代表论文的发表时间）