@ARTICLE{martinez06using, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Using Boosting to Prune Bagging Ensembles}, journal = {Pattern Recognition Letters}, pages = {156--165}, volume = {28}, year = {2007}, number = {1}, abstract = {Boosting is used to determine the order in which classifiers are aggregated in a bagging ensemble. Early stopping in the aggregation of the classifiers in the ordered bagging ensemble allows the identification of subensembles that require less memory for storage, classify faster and can improve the generalization accuracy of the original bagging ensemble. In all the classification problems investigated pruned ensembles with 20% of the original classifiers show statistically significant improvements over bagging. In problems where boosting is superior to bagging, these improvements are not sufficient to reach the accuracy of the corresponding boosting ensembles. However, ensemble pruning preserves the performance of bagging in noisy classification tasks, where boosting often has larger generalization errors. Therefore, pruned bagging should generally be preferred to complete bagging and, if no information about the level of noise is available, it is a robust alternative to AdaBoost.} }@ARTICLE{martinez05switching, author = {Martinez-Munoz, G. and Suarez, A.}, title = {Switching class labels to generate classification ensembles}, journal = {PATTERN RECOGNITION}, year = {2005}, volume = {38}, pages = {1483--1494}, number = {10}, abstract = {Ensembles that combine the decisions of classifiers generated by using perturbed versions of the training set where the classes of the training examples are randomly switched can produce a significant error reduction, provided that large numbers of units and high class switching rates are used. The classifiers generated by this procedure have statistically uncorrelated errors in the training set. Hence, the ensembles they form exhibit a similar dependence of the training error on ensemble size, independently of the classification problem. In particular, for binary classification problems, the classification performance of the ensemble on the training data can be analysed in terms of a Bernoulli process. Experiments on several UCI datasets demonstrate the improvements in classification accuracy that can be obtained using these class-switching ensembles. (c) 2005 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved.}, }@ARTICLE{martinez04using, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Using All Data to Generate Decision Tree Ensembles}, journal = {IEEE Transactions on Systems, Man and Cybernetics part {C}}, year = {2004}, volume = {34}, pages = {393--397}, number = {4} }@INPROCEEDINGS{garcia06evaluation, author = {Sergio Garc\'{\i}a-Moratilla and Gonzalo Martinez-Munoz and Alberto Suarez}, title = {Evaluation of Decision Tree Pruning with Subadditive Penalties}, booktitle = {Intelligent Data Engineering and Automated Learning - IDEAL 2006}, pages = {995--1002}, year = {2006}, abstract = {Recent work on decision tree pruning[1] has brought to the attention of the machine learning community the fact that, in classification problems, the use of subadditive penalties in cost-complexity pruning has a stronger theoretical basis than the usual additive penalty terms. We implement cost-complexity pruning algorithms with general size-dependent penalties to confirm the results of[1] . Namely, that the family of pruned subtrees selected by pruning with a subadditive penalty of increasing strength is a subset of the family selected using additive penalties. Consequently, this family of pruned trees is unique, it is nested and it can be computed efficiently. However, in spite of the better theoretical grounding of cost-complexity pruning with subadditive penalties, we found no systematic improvements in the generalization performance of the final classification tree selected by cross-validation using subadditive penalties instead of the commonly used additive ones.} }@INPROCEEDINGS{martinez06building, author = {Gonzalo Mart\'{\i}nez-Mu\~noz and Aitor S\'anchez-Mart\'{\i}nez and Daniel Hern\'andez-Lobato and Alberto Su\'arez} title = {Building Ensembles of Neural Networks with Class-Switching}, booktitle = {Artificial Neural Networks - ICANN 2006}, pages = {178--187}, year = {2006} }@INPROCEEDINGS{martinez06pruning, author = {Gonzalo Mart\'{\i}nez-Mu\~{n}oz and Alberto Su\'{a}rez}, title = {Pruning in ordered bagging ensembles}, booktitle = {ICML '06: Proceedings of the 23rd international conference on Machine learning}, year = {2006}, isbn = {1-59593-383-2}, pages = {609--616}, location = {Pittsburgh, Pennsylvania}, doi = {http://doi.acm.org/10.1145/1143844.1143921}, publisher = {ACM Press}, address = {New York, NY, USA}, abstract = {We present a novel ensemble pruning method based on reordering the classifiers obtained from bagging and then selecting a subset for aggregation. Ordering the classifiers generated in bagging makes it possible to build subensembles of increasing size by including first those classifiers that are expected to perform best when aggregated. Ensemble pruning is achieved by halting the aggregation process before all the classifiers generated are included into the ensemble. Pruned subensembles containing between 15% and 30% of the initial pool of classifiers, besides being smaller, improve the generalization performance of the full bagging ensemble in the classification problems investigated.} }@inproceedings{hernandez06pruning, title = {Pruning in Ordered Regression Bagging Ensembles}, author = {Hern\'andez-Lobato, D.; Mart\'{\i}nez-Mu\~noz, G.; Su\'arez, A.}, booktitle = {Neural Networks, 2006. IJCNN '06. International Joint Conference on}, year = {2006}, pages = {1266--1273}, abstract = {An efficient procedure for pruning regression ensembles is introduced. Starting from a bagging ensemble, pruning proceeds by ordering the regressors in the original ensemble and then selecting a subset for aggregation. Ensembles of increasing size are built by including first the regressors that perform best when aggregated. This strategy gives an approximate solution to the problem of extracting from the original ensemble the minimum error subensemble, which we prove to be NP-hard. Experiments show that pruned ensembles with only 20% of the initial regressors achieve better generalization accuracies than the complete bagging ensembles. The performance of pruned ensembles is analyzed by means of the bias-variance decomposition of the error.} }@INPROCEEDINGS{martinez05comites, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Comit\'es de \'arboles {IGP}}, booktitle = {Actas del I simposio de inteligencia computacional}, year = {2005}, pages = {277--283}, publisher = {Thomson Press}, }@INPROCEEDINGS{martinez04aggregation, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Aggregation Ordering in Bagging}, booktitle = {Proc. of the {IASTED} International Conference on Artificial Intelligence and Applications}, year = {2004}, pages = {258--263}, publisher = {Acta Press} }@INPROCEEDINGS{martinez02using, author = {Gonzalo Mart{\'{\i}}nez-Mu{\~n}oz and Alberto Su{\'a}rez}, title = {Using All Data to Generate Decision Tree Ensembles}, booktitle = {Proc. of {L}earning'02}, year = {2002}, pages = {181--186}, }