利用覆盖算法对数据进行处理,得到论域U的一个划分,定义一种基于覆盖的条件信息熵,由新的条件信息熵定义新的属性重要性,并证明了对于一致决策表,它与代数定义下的重要性是等价的。以新的属性重要性为启发信息设计约简算法,并给出计算新的条件信息熵的算法。实验结果表明该约简算法能快速搜索到最优或次优约简。
Processing data can be partitioned using covering algorithm.In this paper,a new concept of conditional information entropy is put forward,and then the new significance of an attributes is defined based on this entropy.In a consistent decision table,the equivalence between algebraic significance and conditional information entropy significance of attributes is proved.But it is incorrect in the inconsistent decision table.A heuristic algorithm for knowledge reduction is designed.The experimental results show that this algorithm can find the minimal or optimal reduction.