Skip to Main Content
Based on particle filter, a new training algorithm combining the extended Kalman filter (EKF) for neural network is presented. The new algorithm is firstly applied to train diagonal recurrent neural network (DRNN). A method to evaluate the dynamical performance of DRNN is introduced. Network weights of particles are optimized by the resampling algorithm. Simulation results of the nonlinear dynamical identification verify the validity of the new algorithm.