The training of support vector machines (SVM) involves a quadratic programming problem, which is often optimized by a complicated numerical solver. In this paper, we propose a muc...
Abstract—We consider correlated MIMO multiple access channels with block fading, where each block is divided into training and data transmission phases. We find the channel esti...
We describe an algorithm for support vector machines (SVM) that can be parallelized efficiently and scales to very large problems with hundreds of thousands of training vectors. I...
Abstract. Although widely used to reduce error rates of difficult pattern recognition problems, multiple classifier systems are not in widespread use in off-line signature verifica...
Classifier fusion strategies have shown great potential to enhance the performance of pattern recognition systems. There is an agreement among researchers in classifier combination...
Amin Assareh, Mohammad Hassan Moradi, L. Gwenn Vol...