基于核学习方法的短时交通流量预测

Short-term traffic flow forecasting based on kernel learning methods

基于核学习的强大非线性映射性能,针对短时交通流量预测,提出一类基于核学习方法的预测模型。核递推最小二乘(KRLS)基于近似线性依赖(approximate linear dependence,ALD)技术可降低计算复杂度及存储量,是一种在线核学习方法,适用于较大规模数据集的学习;核偏最小二乘(KPLS)方法将输入变量投影在潜在变量上,利用输入与输出变量之间的协方差信息提取潜在特征;核极限学习机(KELM)方法用核函数表示未知的隐含层非线性特征映射,通过正则化最小二乘算法计算网络的输出权值,能以极快的学习速度获得良好的推广性。为验证所提方法的有效性,将KELM、KPLS、ALD-KRLS用于不同实测交通流数据中,在同等条件下,与现有方法进行比较。实验结果表明,不同核学习方法的预测精度和训练速度均有所提高,体现了核学习方法在短时交通流量预测中的应用潜力。

Based on the powerful nonlinear mapping ability of kernel learning,this paper proposed a class of kernel learning method for the short-term traffic flow forecasting.Kernel recursive least squares(KRLS)method using approximate linear dependence(ALD)technique could reduce the computational complexity and storage capacity,the KRLS method was an online kernel learning method and was suitable for training on large-scale data sets.Kernel partial least square(KPLS)method utilized the covariance between input and output variables to extract latent features.Kernel extreme learning machine(KELM)method used the kernel function to substitute for the unknown nonlinear feature mapping of the hidden layer,in addition,the output weights of the networks could also be analytically determined by using regularization least square algorithm,hence KELM method provided better generalization performance at a much faster learning speed.In order to verify the validity of the proposed kernel learning method,this paper respectively applied the employed KELM,KPLS and ALD-KRLS methods to different traffic flow forecasting instances in different area,compared to the other methods under the same conditions.Experimental results show that the proposed kernel-based learning methods have higher forecasting accuracy and improves training speed in the short-term traffic flow forecasting.