基于信息熵的不完备信息系统属性约简算法

Information entropy based attribute reduction algorithm in incomplete information systems

在现实生活中信息的不完备现象广泛存在,等价关系不一定成立,限制了经典Rough集理论在一些实际问题中的应用。从信息论角度出发引入信息熵的概念,提出了不完备信息系统中知识熵的度量方法,定义了信息观下不完备信息系统知识约简的方法,分析并讨论了与现有其他约简定义的关系,给出了一种基于信息熵的不完备信息系统属性约简算法,通过仿真实验说明了该算法对于不完备信息表知识约简是可行的。

The incompleteness of information exists widely in real life, and the equivalence relation may not be tenable that confines the application of classical rough set theory in practice. Hence, information entropy was introduced in the view of information theory, and the measure method for knowledge entropy in incomplete information systems was proposed. The definition of information based knowledge reduction of incomplete information system was presented, and its relationship with other existing reduction definitions was discussed. An information entropy based attribute reduction algorithm was presented, and simulation results show that this algorithm is effective for knowledge reduction in incomplete information system.