Design of effective neural network ensembles for image classification purposes

Giacinto, Giorgio; Roli, Fabio

doi:10.1016/s0262-8856(01)00045-2

Cited by 360 publications

(196 citation statements)

References 22 publications

Supporting

Mentioning

189

Contrasting

Unclassified

Order By: Relevance

“…Giacinto et al [10] states that classifiers in an ensemble need to be ''accurate and diverse''. Several studies focused on understanding how diversity was handled on various ensemble creation techniques like AdaBoost or Bagging [11,12].…”

Section: Diversity In Ensembles Of Classifiersmentioning

confidence: 99%

“…Classifiers that tend to recognize the same objects correctly will have positive values of Q, and those which commit errors on different objects will render Q negative. Then other similar diversity measures have been proposed, among them, the disagreement measure [21] corresponding to the total proportion of examples for which the two classifiers disagree, the double fault measure [10] which counts the proportion of examples misclassified by both classifiers.…”

Section: Diversity Measuresmentioning

confidence: 99%

“…The diversity measures introduced here above have been extensively used in many applications, particularly in the context of classifier selection from a large set of classifiers [10]. However, is has been observed in practice that explicitly maximizing diversity measures is not as successful as expected.…”

Section: Limits Of Diversity Measuresmentioning

confidence: 99%

See 2 more Smart Citations

Information theoretic combination of pattern classifiers

Meynet¹,

Thiran

2010

Pattern Recognition

View full text Add to dashboard Cite

Section: Diversity In Ensembles Of Classifiersmentioning

confidence: 99%

Section: Diversity Measuresmentioning

confidence: 99%

See 1 more Smart Citation

Information theoretic combination of pattern classifiers

Meynet¹,

Thiran

2010

Pattern Recognition

View full text Add to dashboard Cite

“…It is also possible to create a neural network ensemble using a combination of the above strategies [5]. In general, there are two designs of ANN ensemble methods [6]:…”

Section: Introductionmentioning

confidence: 99%

Neural Networks Fusion for Regression Problems

Shamsoddini¹,

Trinder²

2012

IJMLC

View full text Add to dashboard Cite

Abstract-A common solution to improving the generalization problem and increasing the efficiency of different ANNs is to use ANN ensembles. These methods focus on the possibility of generating different neural nets for a dataset and combining the results for acquiring a more accurate regression. In this paper, a new ensemble method called machine learner fusion-regression (MLF-R) is proposed to increase the accuracy of the results through focusing on difficult samples. The architecture of MLF-R includes two different parts: the first is a training phase from which final nets are selected after a filtering process; the second part is a weighted decision maker including a backpropagation structure which fuses the different nets derived from the first step to predict the outputs. The results demonstrate MLF-R is more efficient than bagging, different boosting methods and the implementation of single ANN methods with 18% to 51% higher accuracy. Moreover, MLF-R offers more stable results compared to the other methods which have been tested in this paper.

show abstract

“…There are effective processes for estimating the accuracy of component classifiers, however, measuring diversity is not easy since there is no generally accepted formal definition. During the past decade, many diversity measures have been designed; to name a few, the Q-statistics [11], the disagreement [9], the double-fault [7], the κ-statistic [5], etc. However, it has been disclosed that existing diversity measures are suspect [10].…”

Section: Introductionmentioning

confidence: 99%

Multi-information Ensemble Diversity

Zhou¹,

Li²

2010

Multiple Classifier Systems

View full text Add to dashboard Cite

Abstract. Understanding ensemble diversity is one of the most important fundamental issues in ensemble learning. Inspired by a recent work trying to explain ensemble diversity from the information theoretic perspective, in this paper we study the ensemble diversity from the view of multi-information. We show that from this view, the ensemble diversity can be decomposed over the component classifiers constituting the ensemble. Based on this formulation, an approximation is given for estimating the diversity in practice. Experimental results show that our formulation and approximation are promising.

show abstract

Design of effective neural network ensembles for image classification purposes

Cited by 360 publications

References 22 publications

Information theoretic combination of pattern classifiers

Information theoretic combination of pattern classifiers

Neural Networks Fusion for Regression Problems

Multi-information Ensemble Diversity

Contact Info

Product

Resources

About