A dynamic classifier ensemble selection approach for noise data

A dynamic classifier ensemble selection approach for noise data
复制标题

噪声数据的动态分类器集成选择方法

DOI:
10.1016/j.ins.2010.05.021
复制
发表时间:
2010-09
影响因子:
8.1
通讯作者:
He, Changzheng
He, Changzheng
中科院分区:
计算机科学1区
文献类型:
--
作者:
Liu, Dunhu;Jiang, Xiaoyi;Xiao, Jin;He, Changzheng

文献摘要

参考文献

被引文献

相似文献

动态分类器集成选择(DCES)在多分类器系统领域具有重要的战略意义。待分类的真实数据往往包含大量的噪声,因此研究各种DCES策略的抗噪声能力非常重要。本文将数据处理分组方法(GMDH)引入到DCES中,提出了一种新的动态分类器集成选择策略GDES-AD。它考虑了集合选择过程中的准确性和多样性。我们在30个UCI数据集上实验测试了GDES-AD和其他六种集成策略,包括三种情况:数据集不包含人工噪声、包含类噪声和包含属性噪声。统计分析结果表明,GDES-AD策略具有较强的抗噪声能力。另外,我们发现Random Subspace比Bagging更适合GDES-AD。此外,对各种策略的分类误差进行偏差-方差分解实验表明,GDES-AD抗噪声能力较强的主要原因是它能较好地降低分类误差中的偏差。
Dynamic classifier ensemble selection (DCES) plays a strategic role in the field of multiple classifier systems. The real data to be classified often include a large amount of noise, so it is important to study the noise-immunity ability of various DCES strategies. This paper introduces a group method of data handling (GMDH) to DCES, and proposes a novel dynamic classifier ensemble selection strategy GDES-AD. It considers both accuracy and diversity in the process of ensemble selection. We experimentally test GDES-AD and six other ensemble strategies over 30 UCI data sets in three cases: the data sets do not include artificial noise, include class noise, and include attribute noise. Statistical analysis results show that GDES-AD has stronger noise-immunity ability than other strategies. In addition, we find out that Random Subspace is more suitable for GDES-AD compared with Bagging. Further, the bias–variance decomposition experiments for the classification errors of various strategies show that the stronger noise-immunity ability of GDES-AD is mainly due to the fact that it can reduce the bias in classification error better.
DOI: 10.1007/3-540-44938-8_31
发表时间: 2003-06
期刊: --
影响因子: --
作者:
Robert E. Banfield;L. Hall;K. Bowyer;W. Kegelmeyer
通讯作者: Robert E. Banfield;L. Hall;K. Bowyer;W. Kegelmeyer
DOI: 10.1109/sbrn.2006.11
发表时间: 2006-10
期刊: 2006 Ninth Brazilian Symposium on Neural Networks (SBRN'06)
影响因子: --
作者:
D. Fagundes;A. Canuto
通讯作者: D. Fagundes;A. Canuto
DOI: --
发表时间: 2000-07
期刊: --
影响因子: --
作者:
Pedro M. Domingos
通讯作者: Pedro M. Domingos
DOI: 10.1007/978-1-4612-4380-9_16
发表时间: 1945-12
期刊: Biometrics
影响因子: 1.9
作者:
F. Wilcoxon
通讯作者: F. Wilcoxon
DOI: 10.1016/j.inffus.2004.04.002
发表时间: 2004-05
期刊: Inf. Fusion
影响因子: --
作者:
T. Windeatt
通讯作者: T. Windeatt