摘要
对目前已有的信息系统属性约简的启发式算法进行了分析,认为它们的时间复杂度不太理想。以快速缩小搜索空间为目的,先利用划分U/A对信息系统进行化简。设计了一个新的度量属性重要性的计算公式,并给出了该公式的递归计算方法。然后给出了一个时间复杂度为max{O(A|U|),O(O|2|U/A|)}的快速属性约简算法,最后,实例及实验结果表明了该算法的有效性。
The heuristic algorithm for attribute reduction in the information system that has arisen at present is analyzed. It＇s time complexity is not good. For reducing searching space as quickly as possible, partition U/A is used to simplify information system. Furthermore, a new formula for measuring the importance of attribution is designed, and the recursive calculating method of the formula is proposed. Then a quick algorithm for attribution reduction is provided. The time complexity of the provided algorithm is max{D（｜A｜｜U｜）,O（｜A｜^2｜U/A｜）}. At last, the efficiency of the new algorithm is illustrated by an example and experiment.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第18期4804-4806,共3页
Computer Engineering and Design
基金
湛江师范学院科研基金项目(L0602)
关键词
信息系统
属性重要性
属性约简
算法复杂度
information system
attribute importance
attribute reduction
algorithm complexity
