对象集变化置信优势关系粗糙集近似集动态更新方法

Incremental Updating Approximations in Confidential Dominance Relation Based Rough Set Under the Variation of the Object Set

对象集变化会造成近似集的改变,动态更新方法可以高效计算近似集.对于不完备有序信息系统,讨论了单个对象增加或删除造成的决策类、置信优势类及置信劣势类的变化情况,给出了基于置信优势关系粗糙集的近似集动态更新方法.通过M atlab在UCI数据集上的实验结果表明,与非增量式方法相比,所提出的动态更新方法在对象集变化情况下计算近似集时间效率更高.

Incremental updating is feasible and effective to update approximations which may evolve under the variation of the object set. When one object added or deleted in incomplete ordered information system, the principles of updating decision classes, confiden- tial dominating and dominated classes are discussed. Then approaches to incremental updating approximations in confidential domi- nance relation rough set under the variation of the object set are proposed. The experiments on UCI datasets developed on Matlab are designed to evaluate the performance of the proposed incremental updating method and non-incremental updating method. The results demonstrate the proposed approaches are effective and efficient under the variation of the object set.