Abstract: Support vector regression is a promising method for the forecast of traffic volume because it uses a risk function consisting of the empirical error and a regularized term which is based on the structural risk minimization principle. This study applies SVR to forecast the traffic volume. Model selection is a key factor for forecast capability. This paper uses spatio-temporal predition model and selects parameters by minimizing LOO upper bound through BFGS varible metric algorithm. Experimental results show that applying this method in traffic volume forecast is feasible and it provides a promising alternative to traffic volume prediction.

Key words: Traffic volume, Forecast, Support vector regression, Leave-one-out(LOO)

摘要： 支持向量回归方法作为以结构风险最小化原理为理论基础的学习算法，可应用于交通量的预测，其中模型参数的选择是预测性能的关键因素。针对交通量的影响因素，应用时空挖掘联合预测模型，给出了BFGS变尺度算法求LOO误差上界最小化值定参的方法，试验证明，该方法对交通量的模型参数选择和预测是有效可行的。

关键词: 交通量, 预测, 支持向量回归, 留一法