标题：Multi-objective Genetic Algorithm Based Selective Neural Networks Ensemble for Concentration Estimation of Indoor Air Pollutants Using Electronic Nose
期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2013
卷号：10
期号：3
出版社：IJCSI Press
摘要：Neural networks ensemble or committee of neural networks is a learning approach where many neural networks are combined to solve a given problem. This approach has been proved to improve the generalization performance of individual networks (base networks), provided these networks are accurate enough while being error-independent (diverse). In this paper, variance inflation factor (VIF) is defined as diversity measure. A multi-objective genetic algorithm (MOGA) with two objectives (ensemble error and the new diversity metric) is used to select appropriate members of the ensemble from a pool of trained neural networks. The proposed method herein called MOGASEN(Multi Objective Genetic Algorithm based Selective ensemble) and other popular ensemble approaches were evaluated on data from an electronic nose (E-Nose) for concentration estimation of four indoor air pollutants (formaldehyde, benzene, toluene, and carbon monoxide). Empirical results show that the proposed method, while having higher capability in reducing the size of the ensemble, was, in most cases, able to outperform other methods.
