基于粗糙集理论文本分类规则的优化方法_王海涌.pdf

20 2 V l. 20 N . 2 2008 6 J urnal f Gan su Sciences Jun . 2008 1, 2 1, 2 王海涌 , 郑丽英 ( 1. , 730070; 2. , 730070) : 文本分类规则的提取和优化是衡量文本分类系统适应性能和分类能力的主要指标. 在基 于粗糙 和模糊聚类理论的文本分类系统中, 结合粗 理论及不完备信息系统理论, 提出了分类规 则的提取和优化方法, 通过实验和分析, 产生用户满意的约简规则, 从而能够快捷迅速地指导新文本 的分类, 提高系统的适应性能和分类能力. : 粗糙 ; 不完备信息系统; 文本分类; 规则约简 : TP391. 41 : A : 1004-0366( 2008) 01-0099-04 An Approach to Extracting Text Classification Decision Rules Based on Rough Set 1, 2 1, 2 WANG Ha-i y ng , ZHENG L-i y ing ( 1 S chool of E lectronic and I nf ormation Engineering , Lanz hou J iaotong University , L anz hou 730070, China ; 2. K ey L a oratory of Op to-Electronic T echnology and I ntelligent Control, M inis try of Education L anz hou J iaotong Univ ers ity , Lanz ho u 730070, China) Abstract: T he extracti n and ptimizati n f T ext classificati n rules are the main indicat rs t evaluate the ability t adapt t and classify a Text classificati n system. Based n r ugh sets and fuzzy clustering the ry f text classificati n system, w ith the help f the the ry f inc mplete inf rmati n system, the pa- per presents the meth d t extract and ptimize a classificati n rule. After experimenting and analy sis, satisfying reducti n rules are btained t help users quickly classify the new text, thus enhancing the abil-i ty t adapt t and classify the system. Key words: r ugh set; inc mplete inf rmati n system ; text classificati n; rule reducti n ,