Rough-set-based Construct of Combined Classifier
Sui Ren-wu · Journal of Hengyang Normal University · 2006
Rough set theory and neural networks are two widely used and complementary tools for uncertain compu-ting.After the discussion of the two theories,a novel method of construct multiple neural network classifiers is pro-posed.The novel method firstly produces many neural network classifiers based on the reduction of data set,allthese classifiers are correlative to dataset but independent to each other;then,these classifiers are combined by theweight determined based on attribute i mportance.Experi ment showthat the proposed method is efficient.