Castelo-Fernández, CésarDe Rezende, Pedro J.Falcão, Alexandre X.Papa, João Paulo [UNESP]2014-05-272014-05-272010-12-15Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), v. 6419 LNCS, p. 467-475.0302-97431611-3349http://hdl.handle.net/11449/72224In this work, a new approach for supervised pattern recognition is presented which improves the learning algorithm of the Optimum-Path Forest classifier (OPF), centered on detection and elimination of outliers in the training set. Identification of outliers is based on a penalty computed for each sample in the training set from the corresponding number of imputable false positive and false negative classification of samples. This approach enhances the accuracy of OPF while still gaining in classification time, at the expense of a slight increase in training time. © 2010 Springer-Verlag.467-475engLearning AlgorithmOptimum-Path Forest ClassifierOutlier DetectionSupervised ClassificationClassification timeFalse negativesFalse positiveForest classifiersLarge datasetsNew approachesSupervised classificationSupervised classifiersSupervised pattern recognitionTraining setsTraining timeClassification (of information)ClassifiersComputer visionData miningLearning algorithmsTrabalho apresentado em evento10.1007/978-3-642-16687-7_62Acesso aberto2-s2.0-786499783759039182932747194