Feature selection based on mutual information and redundancy-synergy coefficient  

作　　者：杨胜 顾钧 

机构地区：[1]InstituteoflmageProcessing＆PatternRecognition,ShanghaiJiaotongUniversity,Shanghai200030,China [2]DepartmentofComputerScience,HongkongUniversityofScienceandTechnology,Hongkong,China

出　　处：《Journal of Zhejiang University Science》2004年第11期1382-1391,共10页浙江大学学报（自然科学英文版）

基　　金：Project supported by the National Natural Science Foundation ofChina (No. 60075007) and the National Basic Research Program(973) of China (No. G1998030401)

摘　　要：Mutual information is an important information measure for feature subset. In this paper, a hashing mechanism is proposed to calculate the mutual information on the feature subset. Redundancy-synergy coefficient, a novel redundancy and synergy measure of features to express the class feature, is defined by mutual information. The information maximization rule was applied to derive the heuristic feature subset selection method based on mutual information and redundancy-synergy coefficient. Our experiment results showed the good performance of the new feature selection method.Mutual information is an important information measure for feature subset. In this paper, a hashing mechanism is proposed to calculate the mutual information on the feature subset. Redundancy-synergy coefficient, a novel redundancy and synergy measure of features to express the class feature, is defined by mutual information. The information maximi- zation rule was applied to derive the heuristic feature subset selection method based on mutual information and redun- dancy-synergy coefficient. Our experiment results showed the good performance of the new feature selection method.

关 键 词：Mutual information Feature selection Machine learning Data mining 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]