属性约简算法CARRDG的改进及其实现技术研究-Core.PDF

160 ， （ ） 计算机工程与应用 2008 44 36 Computer Engineering and Applications 属性约简算法CARRDG 的改进及其实现技术研究 ， 1 2 1 3 陈可赢 ，曾文华 ，施明辉 1，2 1 3 CHEN Ke-ying ，ZENG Wen-hua ，SHI Ming-hui 1.厦门大学软件学院，福建 厦门 361005 2.厦门华厦学院，福建厦门 361005 3.厦门大学智能科学与技术系，福建 厦门 361005 1.School of Software ，Xiamen University ，Xiamen ，Fujian 361005 ，China 2.Xianmen Huaxia College ，Xiamen ，Fujian 361005 ，China 3.Department of Cognition Science ，Xiamen University ，Xiamen ，Fujian 361005 ，China E-mail ：smh@ CHEN Ke-ying，ZENG Wen-hua ，SHI Ming-hui.Research on technologies for improvement and implementation of at－ tribute reduct algorithm CARRDG.Computer Engineering and Applications ，2008 ，44 （36）：160-163. Abstract ：The attribute reduct algorithm CARRDG ，which can efficiently compute total attribute reducts for large scale informa－ tion system ，has been proposed recently.Although it has rigorous theoretical foundation ，further improvement may be possible ，espe－ cially in the implementation level.In addition to the existed three heuristic deep -first searching principles （Member Executive Principle MEP ，Friend Persuade Principle FPP ，Stranger Enter Principle SEP）based on reduct discernibility graph ，a mew heuris－ tic searching principle———Blocking Layer Block Principle （BLBP）has been proposed in this paper to improve the searching effi－ ciency.Since the reasonable data structures have been developed ，BLBP will not increase the implementing complexity of the algo－ rithm.In contrast ，the experimental results by using UCI data show that BLBP exceeds MEP and FPP in trimming efficiency for some large information systems. Key words ：attribute reduct ；reduct discernibility graph ；data mining ；knowledge discovery 摘 要：属性约简算法CARRDG 是近来提出的能计算大型信息系统中所有属性约简的高效算法。针对属性约简算法CARRDG 在 实现技术层面上的可改进之处，在原有的三种约简分辨图深度优先搜索原则（成员