基于信息熵与蚁群优化的属性约简算法  

Attribute Reduction Algorithm Based on Information Entropy and Ant Colony Optimization

作　　者：陈颖悦[1] 陈玉明[2] 

机构地区：[1]厦门理工学院经济与管理实验中心,福建厦门361024 [2]厦门理工学院计算机科学与技术系,福建厦门361024

出　　处：《小型微型计算机系统》2015年第3期586-590,共5页Journal of Chinese Computer Systems

基　　金：基于粒计算与异常点挖掘的网络入侵检测研究(61103246)资助

摘　　要：信息熵是粗糙集理论中度量不确定信息的重要工具之一。蚁群优化算法是一种新型的智能计算的方法,具有分布式、正反馈及启发性搜索等优良的性质,并且在优化计算中已得到了很多应用.最小属性约简问题也是一类优化问题,已有的属性约简算法主要采用Pawlak正域度量属性的重要度,而且求最小约简是NP-hard问题.为此,在分析信息熵度量不确定性数据的基础上,定义信息熵属性重要度概念,引入蚁群优化算法,提出基于信息熵与蚁群优化的最小属性约简算法.该算法发挥蚁群优化算法良好的寻优能力,大多数情况下能够找到最小约简.理论分析与实验结果表明该算法是有效可行的.Information entropy is one of important tool to deal with uncertain information in rough set theory. Ant colony optimization is a novel intelligent computing method which shows many promising characters including distributed,positive feedback and heuristic search. It is numerous applied to optimization computation. Minimal attribute reduction question is an optimization computation. Many existing algorithms of attribute reduction are mainly based on attribute importance with the Pawlak positive region. Find the minimal reduct is also a NP-hard problem. Therefore,on the basis of analysis of information entropy measure of uncertainty on the data, attrib- ute importance is defined by the concept of entropy. Furthermore, a new algorithm for attribute reduction based on information entropy and ant colony optimization is proposed. The proposed algorithm has a good ability for finding a minimal attribute reduct. Finally, theo- retical analysis and experimental results show that the algorithm is efficient and feasible.

关 键 词：粗糙集 蚁群优化 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]