Feature reduction method for rough set based on entropy measurement and genetic algorithm
Jin Chen · Zhendong yu chongji · 2009
A feature mining model was set up for rotating machine fault diagnosis,based on data mining.There are two problems in traditional reduction of rough set.One is the non-uniqueness of the best reduction;the other is the long time for redcution calculating.In the paper,a feature reduction algorithm for rough set(RS) was proposed,based on entropy importance and genetic algorithm(GA).A feature mining tool,RMFMiner,was designed and implemented.This algorithm has been employed to reduce the simulated fault features of rotating machinery and UCI data set to prove its effectiveness.