摘要
影响食品安全评价的因素很多,并且这些因素错综复杂,因此,目前众多算法存在评价指标约简效果差、运算效率低下等问题。针对大多数食品安全评价指标约简算法时间复杂度比较高、指标选取差等问题,文中利用粗糙集理论提出了一种新的解决办法。该方法基于相似矩阵概念,利用属性在相似矩阵中出现的频率给出了属性重要性的计算公式,以此作为启发式知识来约简决策表中的冗余属性,并将折半查询的思想运用到了算法中,加快了约简效率,提高了评价指标选取水平。实验结果表明该算法是简单有效的。
There are many factors which influence the food safety evaluation,and these factors are complicated,therefore,at present many algorithm has problems of poor evaluation index reduction effect,low calculation efficiency. For the problems to most of the food safety evaluation index reduction algorithm,such as high time complexity and poor index selection,by using rough set theory propose a new so-lution. This method is based on the concept of similar matrix,by using the frequency appeared in similar matrix,give the calculation for-mula of attribute importance as heuristic knowledge to reduce redundant attributes of decision table,and apply the idea of binary query to the algorithm,speeding up the reduction efficiency,improving the level of the evaluation index selection. The experiment shows this algo-rithm is simple and feasible.
出处
《计算机技术与发展》
2014年第3期147-149,153,共4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
粗糙集
相似矩阵
折半属性约简
食品安全评价
rough set
similar matrix
half attribute reduction
food safety evaluation