基于主成分分析和极限学习机的公路货运量预测研究

Research on Highway Freight Volume Forecasting Based on Principal Component Analysis and Extreme Learning Machine

提出了一种基于主成分分析和极限学习机的计算方法,并用这种方法对货物运输量进行预测。该方法首先利用主成分分析的方法融合多个影响货运量的因素,获取主成分;然后建立极限学习机模型来预测公路货运量;最后计算模型性能指标并与相关模型进行对比。结果表明,基于主成分分析和极限学习机的货物运输量预测方法具有较高的稳定性和预测精度。

This article proposes a predictive model based on PCA and ELM,and try to use it to forecast freight volume.This method firstly fuses multiple influence factors influen-cing freight volume by using the method of principal component analysis and obtains the principal components;Then Extreme learning machine model is established to forecasting the highway freight volume;Finally,it calculates the performance of the model and compares with the relevant model.It turned out that,this freight volume forecasting method based on PCA and ELM is better than single ELM model or BP neural network model in prediction stability and precision.