数据驱动的可变精度粗糙集噪音阈值获取方法-NSFC.PDFVIP

计算机科学2008Voi．35№．II 数据驱动的可变精度粗糙集噪音阈值获取方法¨ 赵彦钧1’2王国胤1’2胡峰1’2 (西南交通大学信息科学与技术学院 成都610031)1 (重庆邮电大学计算机科学与技术研究所 重庆400065)2 摘要可变精度粗糙集理论是经典粗糙集理论的一种扩展理论。它通过引入噪音阚值B，增强了对噪音数据的适 应性。然而噪音阈值p多是人为设定，这要求有一定先验知识。提出一种方法，完成了数据驱动的噪音周值p的自主 式获取。仿真实验结果表明，按照此方法获取的噪音闻值B能够提高可变精度粗糙集理论获取知识的性能。 关键词可变精度粗糙集，噪音阈值，数据驱动，知识获取 to SetModel Data-driven ofVariablePrecisionThresholdinVariablePrecision ApproachAcquisition Rough ZHAO WANG HU Yan-junl·2 Guo-yinl’2Fengl’2 (schoolofInformationScienceand 810031，CKna)1 Technology，SouthwestJiaoTongUniversity，Chengdu (Instituteof Scienceand ofPostsand 400065，CIliIla)2 Computer Technology。ChongqingUniversity Telecommunications，Chongqing A搬耐Variable setisone fclassicalseL variable ofextensionso introducingprecision precisionrough roughThrough noise variable thresholdB，variable setenhancesits tO data．However，the precisionrough abilityadapt precision set needs domain data-drivenis tO threshold口isupbypeople，whichprior knowledge．A approachproposedcomplete the ofvariable thresholdThe resultsshowthatthevariable self-learningacquisition precision B experiment precision thisdata-drivencan the of ofvariable threshold口acquiredby approachimprovecapabilityknowledgeacquisition preci— sion set． rough Variable set，Variable Keywords precisionrough precisionthreshold，Dat