基于Softplus函数的双隐含层BP神经网络的公路客运量预测  

Highway Passenger Traffic Forecast Based on Double Implicit Layer BP Neural Network Based on Softplus Function

作　　者：徐森 崔舒为 XU Sen;CUI Shuwei(School of Mathematics and Physics,Lanzhou Jiaotong University,Lanzhou,Gansu 730070,China)

机构地区：[1]兰州交通大学数理学院,甘肃兰州730070

出　　处：《南华大学学报（自然科学版）》2020年第1期88-92,共5页Journal of University of South China：Science and Technology

摘　　要：为了提高预测模型的精度,提出一种基于Softplus激活函数的双隐含层BP神经网络的预测方法,提高了模型的非线性学习和泛化能力及预测精度,并改善了网络性能。将该方法应用于公路客运量实际预测中进行有效性验证,结果表明该方法对公路客运量有更好的非线性拟合能力和预测准确性。In order to improve the accuracy of the prediction model,a prediction method based on Softplus activation function for double hidden layer BP neural network is proposed,which improves the nonlinear learning and generalization ability and prediction accuracy of the model,and improves the network performance.The method is applied to the actual prediction of highway passenger traffic volume.The results show that the proposed method has better nonlinear fitting ability and prediction accuracy for highway passenger traffic.

关 键 词：Softplus函数 双隐含层 BP神经网络 公路客运量 

分 类 号：U492.413[交通运输工程—交通运输规划与管理]