摘要
食品安全信息表中,由于采集及传输等原因,经常出现数据缺失导致系统不完备,而粗糙集是填补数据的一种有效工具。目前,虽然利用粗糙集的知识来进行缺失数据填补的方法有很多,但很多方法都没有考虑到原有系统所蕴含的决策规则。为了更好地填补缺失数据,文中在粗糙集理论框架下,借助核值概念进行属性重要性度量,通过构造可辨识矩阵,遵循决策规则填补了信息系统中的缺失数据,消除了影响主题分析的噪音数据,将不完备系统修正为完备系统。实例分析结果表明该算法是有效可行的。
Because of some factors, there are many missed data in the food safety information system. To deal with the problem, rough set theory is usually used in the research field. There are many methods which make use of the knowledge of rough set to fill the missing da- ta, but many of them do not take into account the decision-making rules of original system. In order to better fill the missing data, under the framework of rough set theory, use the concept of core value to measure attribute importance and construct discriminating matrix, so that it can fill the data which better follows the rules of decision-making, eliminating the noise data. Experiments show that the algorithm is simple and effective.
出处
《计算机技术与发展》
2014年第4期193-195,共3页
Computer Technology and Development
基金
中国博士后基金项目(2012M520158)
辽宁省百千万人才基金择优资助项目(2012921058)
辽宁省教育科研项目(L2012397)
辽宁省社科联2014年度辽宁经济社会发展立项课题(2014LSLKTDGLX-02)
关键词
核值
极大完备子系统
可辨识矩阵
core values
huge self-contained subsystem
discriminating matrix
