
Este trabalho propõe uma nova técnica para a maximização de margem, através do tratamento de dados. As várias abordagens de tratamento de dados serão utilizadas no treinamento do Perceptron Espanhol. As distâncias entre os vetores de suporte e a superfície de separação foram utilizadas como métrica para a avaliação da máxima margem. Os vetores de suporte são fornecidos através da Support Vector Machine (SVM), após treinamento com a base de dados sem tratamento. A base utilizada é sintética, linearmente separável, com duas classes, aleatória de distribuição normal. Os resultados sugerem que a nova técnica proporciona uma melhoria na margem gerada com o perceptron espanhol, quando o erro de treinamento é maior.Palavras-chave: Maximização de margens. Perceptron Espanhol. SVM.Neural networks in the data processing for maximizingAbstractThis paper proposes a new technique for maximizing margin, through the processing of data. The various data processing approaches will be used in Spanish Perceptron training. The distances between the support vectors and the separation surface was used as a metric to evaluate the maximum margin. The support vectors are provided by the Support Vector Machine (SVM) after training database without treatment. The base used is synthetic linearly separable, with two classes, random normal distribution. The results suggest that the new technique provides an improvement in margin generated with Spanish Perceptron, when the training error is higher.Keywords: Maximizing margins. Spanish Perceptron. SVM.
| selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | 0 | |
| popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network. | Average | |
| influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically). | Average | |
| impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network. | Average |
