【摘要】 传统的BP神经网络训练算法,导致训练时间长且易于陷入局部极小点.本文将粒子群优化算法用于神经网络预测模型的学习训练.实验结果表明,基于粒子群优化的神经网络学习算法更易于实现,且能更快地收敛于全局最优解.更多还原

【Abstract】 The training of conventional back-propagation method with shortcoming of spending long time and easily plunging into the local minimum.A training algorithm for neural network based on particle swarm optimization(PSO) was investigated.The results show that the training algorithm based on PSO is easily implemented and come to global optimization quickly.更多还原