摘要： 基于粗糙集理论提出一种新的属性重要度的度量方法，引入决策强度的概念，克服经典粗糙集理论约简定义的不完备性及无法获得最优属性约简的缺陷，改进基于信息熵的启发式属性约简算法，通过对既有线CTCS-2级车站列控中心软件测试平台的测试数据的实证分析，成功获得最优属性约简，发现数据之间的潜在联系及规律，给出决策规则，使决策分析更为高效。

关键词: 粗糙集, 信息熵, 约简, 决策规则

Abstract: In order to overcome the incompleteness of reduction definition in classical rough sets theory and no avail of getting the optimal attribute reduction, a new measure of attribute significance is put forward and a concept of decision power is introduced. Accordingly, an improved heuristic algorithm for attribute reduction based on information entropy is proposed. It is analyzed in theory and tested in practice. By means of the test data analysis on CTCS-2 train control center software test platform, the optimal attribute reduction can be successfully obtained by the algorithm, latent relations and rules are also discovered and decision rules are given. It contributes a lot to make more effective decision analysis.

Key words: rough sets, information entropy, reduction, decision rule

中图分类号: 

• TP18