论文部分内容阅读
针对传统神经网络优化算法易陷入局部最优值的问题,在标准粒子群算法的基础上,对粒子速度与位置更新策略进行改进,提出一种基于改进粒子群优化算法的BP神经网络建模方法.使用sinc函数、波士顿住房数据及某钢厂带钢热镀锌生产的实际数据进行验证.结果表明,与标准的反向传播神经网络和支持向量机相比,基于改进粒子群优化的神经网络模型可以有效提高预测精度.
Aiming at the problem that the traditional neural network optimization algorithm is apt to fall into the local optimal value, based on the standard particle swarm optimization algorithm, the particle velocity and position updating strategy is improved. A BP neural network modeling method based on the improved particle swarm optimization algorithm is proposed The results of sinc function, Boston housing data and strip galvanized steel production data show that compared with standard backpropagation neural network and support vector machine, neural network based on improved particle swarm optimization The model can effectively improve the prediction accuracy.