基于信息最大覆盖率蚁群算法的Rough集属性优化约简

摘要

Reduction of redundancy condition attributes is a N-P hard problem which is a kind of time-consuming complex computation.A new approach of reduction of redundancy condition attributes is put forward based on an information consistency relationship of equivalent classification to reduce computation complexity.The best coverage of an information system and the significance of attribute to the system are chosen as a rule of heuristic information.An algorithm of attribute reduction based on ant colony optimization is suggested to improve effectiveness of computation and space complexity.The merit of the optimization of attribute reduction on ant colony algorithm is validated by the proposed application.%属性约简是一个NP难问题,这种算法不仅复杂而且非常耗时。因此本文提出了一种基于等价划分的信息一致性的属性约简算法,降低了属性约简计算复杂度,并将信息最大覆盖率和属性重要性作为蚁群优化算法的信息素进行添加式属性约简,既减小了属性约简的空间复杂度,也提高了计算效率。最后通过一个具体的例子,证明了此算法的有效性和可行性。