Comparison of SVM and ANN performance for handwritten character classification
Fatih Kahraman, Abdulkerim Çapar, Alper Ayvacı, Hasan Demirel, Muhittin Gökmen · 2004
This study is about the selection of classifiers in handwritten character recognition. The aim of the study is to determine the most appropriate classifier type for a given handwritten character feature vector. PCA based features were classified by both multilayer artificial neural networks (ANN) and support vector machines (SVM), and then the recognition results were compared. We selected error backpropagation, resilient backpropagation and scaled conjugate gradients as ANN training methods, while the SVM kernel types selected were linear, RBF and polynomial. The experimental results show that the SVM has better training and test performance than ANN.