再生分辨矩阵与决策熵的不完备决策系统属性约简

为了有效地解决不完备决策系统的属性约简问题,提高属性约简的效率,提出了基于再生分辨矩阵与决策熵的不完备属性约简算法。该算法利用基于容差关系的分辨矩阵来计算相对核,通过再生分辨矩阵计算再生集和再生集属性来缩小加入约简集的条件属性选择范围,以再生集属性的分辨度和决策熵为依据,选择加入约简集的条件属性,并通过实例进行验证分析。结果表明,该算法适用于协调不完备决策系统与不协调不完备决策系统,能够有效地降低时间复杂度,并得到最优属性约简。

Attributes Reduction Based on Regenerative Discernibility Matrix and Decision Entropy in Incomplete Decision System

In order to effectively solve the problem of attributes reduction in incomplete decision system and improve the efficiency of attribute reduction,attribute reduction algorithm based on regenerative discernibility matrix and decision entropy is proposed in incomplete decision system. The algorithm calculates relative core with the dicernibility matrix based on tolerance relation,narrowes the range of condition attribute which join reduction sets by using regenerative discernibilty matrix to calculate regenerative set and its attributes,selectes condition attribute to join reduction sets on the basis of the resolution of regenerative set attributes and decision entropy,and verifies and analysises by an example. Results show that the algorithm is suitable for the coordination and uncoordinated incomplete decision system,can effectively reduce the time complexity and get the optimal attribute reduction.