A new artificial neural network ensemble based on feature selection and class recoding

Sesmero, M. Paz; Alonso-Weber, Juan Manuel; Gutiérrez, Germán; Ledezma, Agapito; Sanchís, Araceli

doi:10.1007/s00521-010-0458-5

Cited by 13 publications

(8 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, the dichotomous classifiers that integrate these models are trained only on partial knowledge and, in some of these architectures (OAA, OAHO), wrong decisions emitted by a binary classifier are not rectifiable [33]. In this scenario, the system accuracy depends mainly on the accuracy of its members but not on their diversity.…”

Section: Ensemble Of Classifiersmentioning

confidence: 99%

“…In [33] we started addressing the resolution of the multiclass classification problems with the proposal of a preliminary framework based on dual base learners. This system was tested on two real problems and the experimental results were rather promising.…”

Section: Binary-complementary Ensemble Architecturementioning

confidence: 99%

“…So, BCE maintains a design in which the base learners are implemented with two coupled classifiers -a binary classifier (B i ) and a complementary classifier (C i )-but now the architecture of the complementary classifier is quite different. In [33] the binary and the complementary classifiers were trained with the whole training set. In the current version of BCE, the complementary classifiers are trained only with the instances that have been labeled as negative for the corresponding binary classifier (see Fig.…”

Section: Binary-complementary Ensemble Architecturementioning

confidence: 99%

“…If the binary classifier is reliable, then these examples will belong to one of the k À 1 classes that are considered to be negative during the learning phase. This step determines that, in contrast with the architecture introduced in [33] where complementary classifiers were trained with examples belonging to all classes, in BCE the i-th complementary classifier is trained only with the instances that have been considered as negative by the i-th binary classifier, i.e., from {X Ài tr }. Similar to in the binary classifier implementation, a process to determine the most relevant feature subset is performed before the complementary classifiers are constructed.…”

Section: Design Of Complementary Classifiersmentioning

confidence: 99%

See 3 more Smart Citations

An ensemble approach of dual base learners for multi-class classification problems

et al. 2015

View full text Add to dashboard Cite

a b s t r a c tIn this work, we formalise and evaluate an ensemble of classifiers that is designed for the resolution of multi-class problems. To achieve a good accuracy rate, the base learners are built with pairwise coupled binary and multi-class classifiers. Moreover, to reduce the computational cost of the ensemble and to improve its performance, these classifiers are trained using a specific attribute subset. This proposal offers the opportunity to capture the advantages provided by binary decomposition methods, by attribute partitioning methods, and by cooperative characteristics associated with a combination of redundant base learners. To analyse the quality of this architecture, its performance has been tested on different domains, and the results have been compared to other well-known classification methods. This experimental evaluation indicates that our model is, in most cases, as accurate as these methods, but it is much more efficient.

show abstract

Section: Ensemble Of Classifiersmentioning

confidence: 99%

Section: Binary-complementary Ensemble Architecturementioning

confidence: 99%

Section: Binary-complementary Ensemble Architecturementioning

confidence: 99%

Section: Design Of Complementary Classifiersmentioning

confidence: 99%

See 2 more Smart Citations

An ensemble approach of dual base learners for multi-class classification problems

et al. 2015

View full text Add to dashboard Cite

show abstract

“…In this case, the new example will be tested by all the classifiers and their output will be combined in order to obtain a final classification. This task can be achieved by using different methods: Linear functions like average function [35]; nonlinear combination methods, like majority voting (Bagging, Boosting) or meta-learning methods [36] [37].…”

Section: ) Combiningmentioning

confidence: 99%

Ensemble method based on individual evolving classifiers

Iglesias

Ledezma

Sanchís

2013

2013 IEEE Conference on Evolving and Adaptive Intelligent Systems (EAIS)

View full text Add to dashboard Cite

Abstract-Humans often seek a second or third opinion about an important matter. Then, a final decision is reached after weighing and combining these opinions. This idea is the base of the ensemble based systems. Ensembles of classifiers are well established as a method for obtaining highly accurate classifiers by combining less accurate ones. On the other hand, evolving classifiers are inspired by the idea of evolve their structure in order to adapt to the changes of the environment.In this paper, we present a proof-of-concept method for constructing an ensemble system based on Evolving Fuzzy Systems. The main contribution of this approach is that the base-classifiers are self-developing (evolving) Fuzzy-rule-based (FRB) classifiers. Thus, we present an ensemble system which is based on evolving classifiers and keeps the properties of the evolving approach classification of streaming data. It is important to clarify that the evolving classifiers are gradually developing but they are not genetic or evolutionary.

show abstract

Generating ensembles of heterogeneous classifiers using Stacked Generalization

Sesmero

Ledezma

Sanchís

2015

WIREs Data Min & Knowl

120

View full text Add to dashboard Cite

Over the last two decades, the machine learning and related communities have conducted numerous studies to improve the performance of a single classifier by combining several classifiers generated from one or more learning algorithms. Bagging and Boosting are the most representative examples of algorithms for generating homogeneous ensembles of classifiers. However, Stacking has become a commonly used technique for generating ensembles of heterogeneous classifiers since Wolpert presented his study entitled Stacked Generalization in 1992. Studies that have addressed the Stacking issue demonstrated that when selecting base learning algorithms for generating classifiers that are members of the ensemble, their learning parameters and the learning algorithm for generating the meta-classifier were critical issues. Most studies on this topic manually select the appropriate combination of base learning algorithms and their learning parameters. However, some other methods use automatic methods to determine good Stacking configurations instead of starting from these strong initial assumptions. In this paper, we describe Stacking and its variants and present several examples of application domains.

show abstract

A new artificial neural network ensemble based on feature selection and class recoding

Cited by 13 publications

References 23 publications

An ensemble approach of dual base learners for multi-class classification problems

An ensemble approach of dual base learners for multi-class classification problems

Ensemble method based on individual evolving classifiers

Generating ensembles of heterogeneous classifiers using Stacked Generalization

Contact Info

Product

Resources

About