Application of Support Vector Machines Techniques in the Neural Network Trainig

Index Terms Neural Network, Backpropagation of Error, Machine Learning, Support Vector, Support Vector Machine Abstract Neural network learning method called Support Vector Neural Trainig was outlined in year 2004 by W. Duch. Goal of this method is to select only those vectors from training set, wich influence at learning process is the biggest. Even though, that this access work well, is possible to make different modifications, so that on a certain data type was the reached accuracy higher or the time necessary to network learning was shorter. This work handle currently these modifications.