决策表知识约简及其条件熵特征刻画

Knowledge Reduction of Decision Tables and Its Characterizations of Conditional Entropy

建立决策表中知识与粗糙熵之间的关系,由此提出决策概念集的条件粗糙熵概念,进而推广为知识的条件熵,并证明知识的条件熵随信息粒度的变小而单调减少的规律,在此基础上给出以不等式为条件的约简判定定理,以此得到知识约简过程中启发式搜索的条件,结合分层递减的思想,设计基于条件熵的决策表知识约简算法。应用实例分析的结果表明,该算法是有效的。

The conditional rough entropy of decision concept sets is proposed followed by establishment of the relation between knowledge and rough entropy in decision tables, and the conditional entropy of knowldege is generalized. The conclusion that conditional entropy of knowledge decreases monotonously as the information granularities become finer is obtained, and the judgment theorem with respect to knowledge reduction is obtained from inequality, and the heuristic method is received. The virtues of hierarchical reduction are carried out to design the heuristic reduction algorithms. Finally, the experimental analyses and comparative results show that it can obtain meaningful and small relative knowledge reduction.