Over-fitting in ensembles of neural network classifiers within ECOC frameworks

被引：0

作者：

Prior, M ^{[1
]}

Windeatt, T ^{[1
]}

机构：

[1] Univ Surrey, CVSSP, Guildford GU2 5XH, Surrey, England

来源：

MULTIPLE CLASSIFIER SYSTEMS | 2005年 / 3541卷

关键词：

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We have investigated the performance of a generalisation error predictor, G(est), in the context of error correcting output coding ensembles based on multi-layer perceptrons. An experimental evaluation on benchmark datasets with added classification noise shows that over-fitting can be detected and a comparison is made with the Q measure of ensemble diversity. Each dichotomy associated with a column of an ECOC code matrix is presented with a bootstrap sample of the training set. G(est) uses the out-of-bootstrap samples to efficiently estimate the mean column error for the independent test set and hence the test error. This estimate can then be used select a suitable complexity for the base classifiers in the ensemble.

引用

页码：286 / 295

页数：10

共 20 条

[1] Learning with ensembles: How over-fitting can be useful
Sollich, P
Krogh, A
[J]. ADVANCES IN NEURAL INFORMATION PROCESSING SYSTEMS 8: PROCEEDINGS OF THE 1995 CONFERENCE, 1996, 8 : 190 - 196
[2] Over-fitting avoidance in probabilistic neural networks
Lotfi, Abdelhadi
Benyettou, Abdelkader
[J]. 2015 WORLD CONGRESS ON INFORMATION TECHNOLOGY AND COMPUTER APPLICATIONS (WCITCA), 2015,
[3] Relation between weight size and degree of over-fitting in neural network regression
Hagiwara, Katsuyuki
Fukunaizu, Kenji
[J]. NEURAL NETWORKS, 2008, 21 (01) : 48 - 58
[4] Deep Convolutional Neural Network Ensembles Using ECOC
Ahmed, Sara Atito Ali
Zor, Cemre
Awais, Muhammad
Yanikoglu, Berrin
Kittler, Josef
[J]. IEEE ACCESS, 2021, 9 : 86083 - 86095
[5] Enhancing classifiers through neural network ensembles
Onaci, Alexandru
Vidrighin, Camelia
Cuibus, Mihai
Potolea, Rodica
[J]. ICCP 2007: IEEE 3RD INTERNATIONAL CONFERENCE ON INTELLIGENT COMPUTER COMMUNICATION AND PROCESSING, PROCEEDINGS, 2007, : 57 - +
[6] SoftTarget Regularization: An effective technique to reduce over-fitting in Neural Networks
Aghajanyan, Armen
[J]. 2017 3RD IEEE INTERNATIONAL CONFERENCE ON CYBERNETICS (CYBCONF), 2017, : 222 - 226
[7] FIGHTING OVER-FITTING WITH QUANTIZATION FOR LEARNING DEEP NEURAL NETWORKS ON NOISY LABELS
Tallec, Gauthier
Yvinec, Edouard
Dapogny, Arnaud
Bailly, Kevin
[J]. 2023 IEEE INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, ICIP, 2023, : 575 - 579
[8] A dynamic method that emphasizes diversity for constructing ensembles of neural network classifiers
Zheng, JJ
Gan, RC
Wang, JX
[J]. PROCEEDINGS OF THE 2005 INTERNATIONAL CONFERENCE ON NEURAL NETWORKS AND BRAIN, VOLS 1-3, 2005, : 763 - 766
[9] Using One-Dimensional Linear Interpolation Method to Check Over-Fitting in Neural Network with Multi-Dimensional Inputs
Hung, Bao-Lien
Hung, Hsin-An
[J]. FRONTIERS OF MANUFACTURING AND DESIGN SCIENCE IV, PTS 1-5, 2014, 496-500 : 2228 - +
[10] Artificial neural networks applied in the analysis of chemical data .1. Approximation for trend and over-fitting
Liu, P
Liang, YZ
Zhang, L
Yu, RQ
[J]. CHEMICAL JOURNAL OF CHINESE UNIVERSITIES-CHINESE, 1996, 17 (06): : 861 - 865

← 1 2 →