最小最大概率回归机在短时交通流预测中的应用

Application of Minimax Probability Machine Regression in Short-term Traffic Flow Prediction

为了提高短时交通流预测精度,更加精确地进行交通流规划和管理,引入一种新颖的基于最小最大概率回归机(MPMR)的短时交通流预测模型。针对北京某公路监测站实测的交通流数据集以及英国某地区实测的交通流数据集,利用基于MPMR的预测模型进行短时交通流预测,并与常规的基于神经网络、基于支持向量机(SVM)以及基于自适应神经模糊推理系统(ANFIS)预测模型的预测性能进行比较。试验结果表明,基于MPMR的短时交通流预测模型可以很好地跟踪实际流量值,在同等条件下相比常规预测模型的预测精度更优,验证了所提出模型的有效性。

To improve the precision of short-term traffic flow prediction and make programming and management of traffic flow more accurately, a novel short-term traffic flow prediction model based on minimax probability machine regression （MPMR） is proposed. The prediction model based on MPMR is applied in the predictions of the short-term traffic flows using measured data set of a Beijing highway monitoring station and traffic flow data set of a region in UK, and its prediction performance is compared with those of prediction models based on neural network, support vector machine （SVM） and adaptive neural fuzzy inference system （ANFIS）. The experiment result shows that the short-term traffic flow prediction model based on MPMR can track the actual traffic flow well, and the accuracy of prediction is better than those of traditional prediction models under the same condition. The effectiveness of the proposed prediction model is verified.