关键词: 文本分类, 特征提取, 互信息算法, 统计量算法

Abstract: Firstly this paper analyzed the importance of applying page categorization in digital library, then it introduced the connation and basic technologies of page categorization. Also the author studied two main feature selection methods specially: Mutual Information algorithm and Chi-square algorithm. Meanwhile, the paper pointed the current weaknesses of the two feature selection methods and put forward corresponding improving measures.

Key words: Page categorization