首页 | 本学科首页   官方微博 | 高级检索  
     

基于互信息的并行约简
引用本文:闫电勋,邓大勇,金皓苹,陈林. 基于互信息的并行约简[J]. 浙江师范大学学报(自然科学版), 2012, 0(3): 300-304
作者姓名:闫电勋  邓大勇  金皓苹  陈林
作者单位:浙江师范大学数理与信息工程学院
摘    要:将并行约简理论和信息论相结合,提出了信息论意义下的并行约简概念(原并行约简理论是代数意义下的并行约简),并利用互信息的概念提出了基于互信息的并行约简算法.信息论意义下的并行约简的优点在于:处理不一致数据时可以保留比代数意义下的并行约简更多的分类信息.最后从理论上探讨了信息论并行约简和代数并行约简的一致性和差异性,将两者进行了有机统一.

关 键 词:粗糙集  决策系统  并行约简  动态约简  互信息

Parallel reduction based on mutual information
YAN Dianxun,DENG Dayong,JIN Haoping,CHEN Lin. Parallel reduction based on mutual information[J]. Journal of Zhejiang Normal University Natural Sciences, 2012, 0(3): 300-304
Authors:YAN Dianxun  DENG Dayong  JIN Haoping  CHEN Lin
Affiliation:(College of Mathematics,Physics and Information Engineering,Zhejiang Normal University,Jinhua Zhejiang 321004,China)
Abstract:It was proposed parallel reduction in information view(the original parallel reduction was in algebraic view) by combining the information theory with the parallel reduction theory,and the corresponding algorithm with the concept of mutual information was also presented.The parallel reduction in information view could preserve more classification information than parallel reducts in algebraic view when used to deal with inconsistent data.Moreover,the consistency and differences of these two parallel reduction were discussed and proposed a unified dealing.
Keywords:rough sets  decision system  parallel reduction  dynamic reduction  mutual information
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号