论文部分内容阅读
在SVM预测模型中引入交叉验证和网格搜索算法,优化惩罚因子和核函数的参数,建立了改进的SVM预测模型,并应用于短时交通流预测进行实证分析。以某城市道路的实时数据来对模型进行验证,预测结果表明了模型的有效性。
In the SVM prediction model, the cross-validation and grid search algorithms are introduced to optimize the parameters of penalty factor and kernel function. An improved SVM prediction model is established and applied to short-term traffic flow prediction for empirical analysis. The model is verified with the real-time data of a city road, and the prediction results show the validity of the model.