基于Lagrange支持向量回归机的短时交通流量预测模型的研究

Short-term Traffic Flow Prediction Model Based on Lagrange Support Vector Regression

在智能交通系统中,进行实时、准确的交通流预测是交通控制和交通流诱导的关键之一,直接影响交通控制和交通诱导的效果。基于支持向量机,提出了一种Lagrange支持向量回归机的交通流量短时预测模型,能够实现对交通流量的有效预测。仿真试验表明,Lagrange支持向量回归机具有良好的泛化性能、更快的迭代速度,预测结果优于改进的BP神经网络。

In intelligent transportation system（ITS）, the accurate prediction of real-time traffic flow is one of keys to traffic control and traffic guidance, and its prediction results will have direct effect on traffic control and traffic guidance. Based on Support Vector Machine （SVM）, a short-term traffic flow prediction model using Lagrange support vector regression （LSVR） was proposed, which could efficiently predict traffic flow. The simulation results show that LSVR has better generalization ability, quicker iterative speed, and better prediction effects than improved BP neural networks.