摘要
通过对信息系统中属性进行多视角的重要度度量,构造一种更为优越的特征选择算法。以粒计算理论为基础,在模糊近似空间中引入模糊粒度,在此基础上提出模糊条件熵的概念,将模糊粒度与模糊条件熵组合作为属性重要度的度量,给出一种基于模糊近似空间组合度量的特征选择算法。实验结果表明,该算法在特征子集和算法效率方面具有较好的优越性。
A more superior feature selection algorithm was constructed by measuring the importance from multiple perspectives for attributes in information system.Based on the theory of granular computing,the fuzzy granularity in the approximation space was introduced.The concept of fuzzy conditional entropy was proposed on that basis,in addition,fuzzy granularity and fuzzy conditional entropy were grouped as the measure of attribute importance.The feature selection algorithm based on the combination measurement of fuzzy approximation space was introduced.Experimental results show that the proposed algorithm is superior in terms of the feature subsets and the algorithm efficiency.
作者
费贤举
刘金硕
田国忠
EI Xian-ju 1, LIU Jin-shuo 2, TIAN Guo-zhong 1(1. School of Computer Information and Engineering, Changzhou Institute of Technology, Changzhou 213032, China;2. School of Computer, Wuhan University, Wuhan 430072, Chin)
出处
《计算机工程与设计》
北大核心
2018年第7期1911-1916,共6页
Computer Engineering and Design
基金
国家自然科学基金项目(61363004)
关键词
特征选择
模糊近似空间
模糊粒度
模糊条件熵
组合度量
feature selection
fuzzy approximation space
fuzzy granularity
fuzzy conditional entropy
composite measurement