基于信息熵的属性约简在多属性决策中的应用  

Application for Multi-Attribute Decision Making Based on Attribute Reduction of Information Entropy

作　　者：杨春林[1] 

机构地区：[1]兰州商学院信息工程学院,甘肃兰州730020

出　　处：《数学的实践与认识》2013年第3期97-102,共6页Mathematics in Practice and Theory

摘　　要：属性约简是在信息系统中的一个重要操作.分类是属性约简的基础,且直接在大数据集上进行属性约简往往存在效率低下的问题.以分类为基础提出了一种基于信息熵的信息系统属性约简算法.算法通过信息熵的计算,在属性约简的同时对原信息系统逐层分解,从而实现了属性的约简并缩小了搜索空间.提出了依据信息熵来确定属性的不必要性及简约属性集,应用在多属性决策中所带来的优势.Attribute Reduction is an important operation in information system. Classifi- cation is foundation of attribute reduction, and always reduce the efficiency if directly carry on attribute reduction on large DS. The essay is based on classification. It puts forward an algorithm which is based on information system attribute reduction of information entropy. The algorithm computes by information entropy, both attribute reduction and layer decompo- sition proceed simultaneously to complete attribute reduction as well reduce the search space. It shows advantages which are brought by the application based on information entropy to define attribute＇s un-necessity and simple attribute set in multi-attribute decision making.

关 键 词：多属性决策 属性约简 信息熵 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]