一种负相关算法结合bagging算法.pdf

Evolutionary Random Neural Ensembles Based on
Negative Correlation Learning
Huanhuan Chen and Xin Yao
Abstract— This paper proposes to incorporate bootstrap of existing ensemble algorithms pay more attention to either
data, random feature subspace and evolutionary algorithm with accuracy or diversity. For example, Bagging, Bagging of fea-
negative correlation learning to automatically design accurate ture and random forests focus on diversity by randomization
and diverse ensembles. The algorithm utilizes both bootstrap
of training data and random feature subspace techniques to of data or/and feature and Adaboost concentrates on accuracy
generate an initial and diverse ensemble and evolves the ensem- by changing the sampling weight of each sample to reduce
ble with negative correlation learning. The idea of generating the training error.
ensemble by simultaneous randomization of data and feature Bagging [7] and Random feature subspace [8], the well-
is to promote the diversity within the ensemble and encourage known ensemble algorithms which have attracted an ex-
different individual NNs in the ensemble to learn different parts
or aspects of the training data so that the ensemble can learn tensive research interest, employ bootstrap sampling [9] of
better the entire training data. Evolving the ensem