【摘要】 准确、高效的交通流预测是实现交通诱导和交通控制的前提和关键。针对传统机器学习方法需要人工构造特征、无法充分提取交通流的时空特征等问题,提出一种混合预测模型,该模型结合卷积神经网络(Convolutional Neural Network,CNN)和XGBoost (Extreme Gradient Boosting)各自的优势,在网络底层使用CNN对交通流数据进行特征的自动提取和选择,并将得到的高维特征向量输入到XGBoost模型中进行预测。为验证模型有效性,取高速路段的交通流数据对CNN模型、XGBoost模型和CNN-XGBoost模型进行实验对比,结果表明,在预测精度上,CNN-XGBoost模型比CNN模型和XGBoost模型分别提高了约6%和7%,是一种有效的短时交通流预测模型。更多还原

【Abstract】 Accurate and efficient traffic forecasting is crucial for traffic guidance and control. Traditional machine learning methods have some shortcomings in this respect, such as artificial structural feature, inefficient feature extraction and so on, so a new hybrid prediction that combine the advantages of convolutional neural network(CNN) and extreme gradient boosting(XGBoost) is proposed. At the bottom of the network, CNN is used to automatically extract and select the features of traffic flow data, and the obtained high-dimensional feature vectors are input into the XGBoost model for prediction. In order to verify the validity of the model, the CNN model, XGBoost model and CNN-XGBoost model are compared with the actual high-speed traffic flow data. The results show that the prediction accuracy of CNN-XGBoost model is improved by 6% and 7% compared with CNN model and XGBoost model respectively, so it is an effective short-term traffic flow prediction model.更多还原