基于区间值信息系统的信息熵增量式属性约简算法

Incremental Attribute Reduction of Information Entropy in Interval valued Information System

文章针对区间值信息系统中对象动态增加的数据环境,提出一种信息熵的区间值信息系统增量式属性约简算法。首先研究了区间值信息系统增加单个对象时,信息系统信息熵度量的增量式更新,理论分析表明这种增量式更新方法具有很高的计算效率;然后以单个对象增加时信息熵度量的增量式更新为基础,通过逐个迭代的方式实现了多个对象增加时信息熵度量的增量式更新;最后提出了一种区间值信息系统信息熵的增量式属性约简算法。实验结果分析表明,所提出的增量式属性约简算法比非增量式算法具有更高效的动态属性约简性能。

An incremental attribute reduction algorithm based on information entropy is proposed for the data environment of objects dynamic increase in interval-valued information system.In this paper,we first study the incremental updating of information entropy measurement when a single object is added to an interval-valued information system.Theoretical analysis shows that the incremental updating method has high computational efficiency.Then,based on the incremental update of information entropy measurement when a single object is added,the incremental update of information entropy measurement when multiple objects are added is realized by iteration one by one.Finally,an incremental attribute reduction algorithm of information entropy of interval-valued information system is proposed.The experimental results show that the incremental attribute reduction algorithm is more efficient than the non incremental algorithm in dynamic attribute reduction.