摘要
神经网络具有结构简单,鲁棒性强,能够逼近任意函数的非线性映射能力,在多个领域得到了广泛应用。但其梯度下降法容易陷入局部最优,训练效率较低。采用粒子群算法(PSO)对BP神经网络进行改进,利用粒子群算法为BP神经网络提供精确的全局搜索能力,提高其训练效率和预测精度。基于建筑物实际沉降观测数据,对BP神经网络和PSO-BP神经网络进行对比分析。结果表明,PSO-BP神经网络的训练效果获得了较大提升,预测精度提升了约61%,预测结果明显优于传统BP神经网络。
Neural Network is widely used in many fields because of its simple structure,strong robustness,and its ability to approximate nonlinear functions of any function. However,the gradient descent method is easy to fall into local optimum and the training efficiency is low. In this paper,the Particle Swarm Optimization(PSO)is used to improve the BP neural network,and the PSO is used to provide accurate global search ability for the BP neural network,so as to improve its training efficiency and prediction accuracy.Based on the actual settlement observation data of buildings,BP neural network and PSO-BP neural network are compared and analyzed. The results show that the training effect of PSO-BP neural network has been greatly improved,and the prediction accuracy has increased by about 61%. The prediction result is obviously better than that of the traditional BP neural network.
出处
《建筑技术开发》
2019年第13期103-104,共2页
Building Technology Development
