一种新的基于决策熵的决策表约简方法

分析了在知识约简过程中经典粗糙集理论决策表知识约简方法的不足.以知识粗糙熵为基础,将一致和不一致对象分开,提出决策熵的概念及其属性重要性,在此基础上给出约简的判定定理;然后以条件属性子集的决策熵来度量其对决策分类的重要性,提出一种新的知识约简启发式方法.理论分析和实验结果表明,基于决策熵的属性重要性是一种更有效的启发式信息,该方法时间复杂度较低,有助于搜索最小或次优约简.

New reduction method based on decision information entropy in decision table

In decision table, the disadvantages of classical rough reduction algorithm were analyzed. Based on the recent rough entropy of knowledge, the new decision information entropy was proposed with separating consistent objects from inconsistent objects, and the new significance of an attribute was defined. The judgment theorem based on this entropy was obtained with respect to knowledge reduction. Condition attributes were considered to estimate the significance for decision classes, and a heuristic algorithm was proposed. Theoretical analysis shows that the proposed heuristic information was better and more efficient than the others, and experimental results prove the validity of the heuristic algorithm in searching the minimal or optimal reduction.