摘要
多种机器学习和深度学习的模型和算法应用于短时交通流量预测,但是,大多数模型尤其是深度学习模型对训练样本的数量要求较高。为此,提出了一种基于数据扩展的短时交通流量预测方法,该方法基于自编码神经网络分别结合长短时记忆神经网络(LSTM)和支持向量机回归(SVR)构建预测模型,该模型利用自编码神经网络扩展的数据分别训练长短时记忆神经网络和支持向量回归进行交通流量的预测,结果表明,所提出的预测模型具有较高的精度和较好的泛化能力。
The variety of models and algorithms of machine learning and deep learning applied to short-term traffic flow forecasting.However,most models,especially deep learning models,requires a high number of training samples.In order to overcome this deficiency,a short-term traffic flow prediction method based on data expansion is proposed.A prediction model is constructed based on Auto-Encoder combined with long-term memory neural network(LSTM)and support vector machine regression(SVR).Long-short memory neural networks and support vector regression are trained based on data extended by Auto-Encoder neural networksto predict traffic flow.The results show that the proposed prediction model has higher accuracy and better generalization ability.
引文
[1] Levin M,Tsao Y D.On forecasting freeway occupancies and volumes(abridgment)[J].Transportation Research Record,1980(773).
[2] Williams B M,Hoel L A.Modeling and Forecasting Vehicular Traffic Flow as a Seasonal ARIMA Process:Theoretical Basis and Empirical Results[J].Journal of Transportation Engineering,2003,129(6):664-672.
[3] Guo J.Huang W,Williams B M.Adaptive Kalman filter approach for stochastic short-term traffic flow rate prediction and uncertainty quantification[J].Transportation Research Part CEmerging Technologies,2014,(43):50-64.
[4] Chen Y,Zhang Y,Hu J.Multi-Dimensional Traffic Flow Time Series Analysis with Self-Organizing Maps[J].Tsinghua Science and Technology,2008,13(2):220-228.
[5]Jiang X,Adeli H.Dynamic wavelet neural network model for traffic flow forecasting[J].Journal of transportation engineering,2005,131(10):771-779.
[6] Castroneto M,Jeong Y S,Jeong M K,et al.Online-SVR for short-term traffic flow prediction under typical and atypical traffic conditions[J].Expert Systems with Applications An International Journal,2009,36(3):6164-6173.
[7] Smith B L,Demetsky M J.Traffic Flow Forecasting:Comparison of Modeling Approaches[J].Journal of Transportation Engineering,1997,123(4),DOI:10.1061/(ASCE)0733-947X(1997)123:4(261).
[8] Bengio,Y.Learning Deep Architectures for AI[J].Foundations and Trends?in Machine Learning,2009,2(1):1-127.
[9] Hinton G E,Salakhutdinov R R.Reducing the dimensionality of data with neural networks[J].science,2006,313(5786):504-507.
[10]Collobert R,Weston J.A unified architecture for natural language processing:Deep neural networks with multitask learning[C]//Proceedings of the 25th international conference on Machine learning.ACM,2008:160-167.
[11]Goodfellow I J,Bulatov Y,Ibarz J,et al.Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks[J].2013,arXiv:1312.6082.
[12]Huval B,Coates A,Ng A.Deep learning for class-generic object detection[J].arXiv preprint:1312.6885,2013.
[13]Shin H C,Orton M R,Collins D J,et al.Stacked autoencoders for unsupervised feature learning and multiple organ detection in a pilot study using 4Dpatient data[J].IEEE transactions on pattern analysis and machine intelligence,2013,35(8):1930-1943.
[14]Lv Y,Duan Y,Kang W,et al.Traffic flow prediction with big data:a deep learning approach[J].IEEE Transactions on Intelligent Transportation Systems,2015,16(2):865-873.
[15]Tian Y,Li P.Predicting Short-Term Traffic Flow by Long Short-Term Memory Recurrent Neural Network[C]//IEEE International Conference on Smart City/socialcom/sustaincom.2015.
[16]Ma X,Tao Z,Wang Y,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J].Transportation Research Part C:Emerging Technologies,2015,54:187-197.
[17]Hochreiter S,Schmidhuber J.Long short-term memory[J].Neural computation,1997,9(8):1735-1780.
[18]易万,罗晶,李勇,等.基于自编码神经网络建立的搜索信息模型[J].计算技术与自动化,2015,(2):117-121.