一种基于Rough集理论的属性约简启发式算法.pdfVIP

26 3 Vol.26 No. 3 2005 3 MINI- MICRO SYS EMS M ar. 2005 Rough 何　明, 冯博琴, 马兆丰, 傅向华 ( , 710049) E-mail: ming he1314@ sina. com - : 属性约简是知识发现中的关键问题之一. 为了能够有效地获取决策表中属性的最小 对约简, 在Rough 集理论的基础 上构造了一 个新的算子, 将信息论角度定义的属性的重要性作为启发式信息, 来描述在决策表中条件属性所提供的知识对决策 属性的影响; 并采用宽度优先搜索策略, 提出了一种新的属性约简启发式算法. 以原始条件属性集为起点并结合算子, 通过向属 性核的递减式逼近, 得到属性的最小 对约简. 实例分析表明, 该算法能有效地对决策表属性进行约简. : Rough ; ; ; : 311: : 1000-1220( 2005) 03-0356-04 P A Heuristic Algorithm for Reduction of Attributes Based on Rough Set Theory , - , - , - HE Ming FENG Bo qin MA Zhao feng FU Xiang hua ( Dep artment of Comp uter Science and Technology of X ian J iaotong University , X ian 710049, China) Abstract: Attribut e reduction is one of t he key problems in the know ledge discovery. Based on the rough set theory, a new op- erat or is constructed t o effect ively achieve the minimal relative reduct ion of attribute in the decision t able Regarding t he signif-. icance of attributes defined from the view point of information theory as heuristic informat ion and applying the breadth-first search strategy a new heuristic algorithm for reduction of attribute is proposed Acquiring optimal relative reduction by de-, . scending approach to core of attribute from original set of conditional attribute and combining w ith operator. Finally, the ex- perimental results show that this algorithm w as effective in attributes reduction of decision tables. Key words: rough set; attribut e reduction; heuristic algorithm; core 1 Ziarko. W NP-hard [9] . - . NP hard [ 1] Rough