A dynamic ensemble learning algorithm for neural networks

被引:424
作者
Alam, Kazi Md Rokibul [1 ]
Siddique, Nazmul [2 ]
Adeli, Hojjat [3 ,4 ,5 ]
机构
[1] Khulna Univ Engn & Technol, Dept Comp Sci & Engn, Khulna 9203, Bangladesh
[2] Ulster Univ, Sch Comp Engn & Intelligent Syst, Derry BT48 7JL, Londonderry, North Ireland
[3] Ohio State Univ, Dept Neurosci, Columbus, OH 43210 USA
[4] Ohio State Univ, Dept Neurol, Columbus, OH 43210 USA
[5] Ohio State Univ, Dept Biomed Informat, Columbus, OH 43210 USA
关键词
Neural network ensemble; Backpropagation algorithm; Negative correlation learning; Constructive algorithms; Pruning algorithms; SELECTION; DIVERSITY; CLASSIFICATION; CLASSIFIERS; COMBINATION; DESIGN;
D O I
10.1007/s00521-019-04359-7
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper presents a novel dynamic ensemble learning (DEL) algorithm for designing ensemble of neural networks (NNs). DEL algorithm determines the size of ensemble, the number of individual NNs employing a constructive strategy, the number of hidden nodes of individual NNs employing a constructive-pruning strategy, and different training samples for individual NN's learning. For diversity, negative correlation learning has been introduced and also variation of training samples has been made for individual NNs that provide better learning from the whole training samples. The major benefits of the proposed DEL compared to existing ensemble algorithms are (1) automatic design of ensemble; (2) maintaining accuracy and diversity of NNs at the same time; and (3) minimum number of parameters to be defined by user. DEL algorithm is applied to a set of real-world classification problems such as the cancer, diabetes, heart disease, thyroid, credit card, glass, gene, horse, letter recognition, mushroom, and soybean datasets. It has been confirmed by experimental results that DEL produces dynamic NN ensembles of appropriate architecture and diversity that demonstrate good generalization ability.
引用
收藏
页码:8675 / 8690
页数:16
相关论文
共 63 条
[1]   Bayesian selective combination of multiple neural networks for improving long-range predictions in nonlinear process modelling [J].
Ahmad, Z ;
Zhang, J .
NEURAL COMPUTING & APPLICATIONS, 2005, 14 (01) :78-87
[2]  
[Anonymous], 2013, UCI MACHINE LEARNING
[3]  
[Anonymous], 2004, FORMAL PATTERN ANAL
[4]  
[Anonymous], 1995, CONNECT SCI
[5]  
Ash T., 1989, Connection Science, V1, P365, DOI 10.1080/09540098908915647
[6]   Dynamic selection of classifiers-A comprehensive review [J].
Britto, Alceu S., Jr. ;
Sabourin, Robert ;
Oliveira, Luiz E. S. .
PATTERN RECOGNITION, 2014, 47 (11) :3665-3680
[7]  
Brown G, 2005, J MACH LEARN RES, V6, P1621
[8]   Modelling of color perception of different eye colors using artificial neural networks [J].
Cengiz, Candan ;
Kose, Erdogan .
NEURAL COMPUTING & APPLICATIONS, 2013, 23 (7-8) :2323-2332
[9]   Semisupervised Negative Correlation Learning [J].
Chen, Huanhuan ;
Jiang, Bingbing ;
Yao, Xin .
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS, 2018, 29 (11) :5366-5379
[10]   Multiobjective Neural Network Ensembles Based on Regularized Negative Correlation Learning [J].
Chen, Huanhuan ;
Yao, Xin .
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING, 2010, 22 (12) :1738-1751