应用支持向量机的交通流状态预测方法研究

支持向量机（support vector machine，sVM）是近年来出现的立足于统计学习理论（statislical learning theory，SLT）的VC维理论和结构风险最小化原则基础上的机器学习方法，在数据挖掘及分类中具有特点和优越性．为了提高交通流状态预测的精度及效率，研究支持向量机应用于数据泛化及分类的方法，并建立模型，在实测数据的基础上进行交通流状态的判定及预测．实验结果表明该方法学习及预测速度快、效率高，并且误差可控，具有较高的精确度（本文中实例精度高于95％），应用前景广泛．

Research on support vector machine-used traffic flow patterns prediction methods

Support vector machine（SVM） is a kind of new machine learning algorithm emerged in recent years based on VC-dimension theory and structure risk minimization principle of statistical learning theory（SLT）. SVM is advanced in the area of data mining and classification. To improve the accuracy and efficiency of predication about the traffic flow pattern, the method about SVM-used data classification is researched, and an appropriate model is built to predict the traffic flow patterns based on real data. Results show that this method has a high efficiency and accuracy which leads a broad application prospect.