基于支持向量机的都市圈客运量预测模型

将都市圈客运量样本数据集分为训练集、测试集和检验集, 采用最小最终误差预测准则确定预测值的损失函数参数与惩罚因子, 选取ε-不敏感损失函数与高斯核函数减小预测复杂性, 构建了基于支持向量机的都市圈客运量预测模型, 并通过逐渐改变损失函数、惩罚因子与高斯核函数参数的取值, 对京津冀都市圈客运量进行了预测。预测结果表明: 客运量预测的平均相对误差为0.15%, 预测值与实测数据拟合良好, 整体变化趋势一致, 反映了预测模型的可靠性。

Prediction model of passenger transport volume in metropolitan region based on support vector machine

The sample data set of metropolitan region's passenger transport volume was divided into training set, testing set and examining set, the loss function's parameters and penalty factor were determined according to the final prediction error criterion, the ε-insensitive loss function and the Gaussian kernel function were chosen to decrease prediction complexity, and a prediction model of metropolitan region's passenger transport volume was proposed based on support vector machine. The passenger transport volume of Beijing-Tianjin-Hebei Metropolitan Region was predicted by gradual changing the parameter values of loss function, penalty factor and Gaussian kernel function. Prediction result shows that the average relative error is 0.15%, and the total changing trends of prediction value and testing value are same, so the model is credible.