A neural network approach for least squares support vector machines learning
Han Liu, Ding Liu · 2009
A new neural network for least squares support vector machines (LS-SVM) learning, which combines LS-SVM with recurrent neural networks, is proposed based on the learning network of standard SVM. It is obtained using Lagrange multipliers directly which eliminates the nonlinear parts of the standard SVM learning network. The proposed network can be used for classification and regression application, whose topology easily adapts to the implementation of analog circuits implementation. The simulation experiment results based on Simulink and Spice illustrate the effectiveness of the proposed neural network.