MEASUREMENT OF WORD RELEVANT DEGREE BASED ON CO-OCCURRENCE INFORMATION

摘要：

The paper presents a new method based on co-occurrence information to estimate word relevant degree. It adopts the special document structure of scientific papers and the position information of words in order to establish the global co-occurrence relationship between words. At the same time, punctuation information is used to establish the local co-occurrence relationship of words. Then we improve the log frequency and global entropy weight method to calculate the relevant degree of words. The experiment results show that our method is effective in measurement of word relevant degree.

关键词： co-occurrence words co-occurrence information word relevance degree the log frequency and global entropy weight method

作者: Dongna Shi Wei Li Cong Wang

作者单位: Information Engineering college,Beijing University of Posts and Telecommunications,Beijing,China

会议类型: 国际会议

会议名称: 2008年拟人系统国际会议(2008 International Conference on Humanized Systems )(ICHS’08)

会议地点: 北京

会议语种:英文

在线出版日期: 2008-10-18（万方平台首次上网日期，不代表论文的发表时间）