基于时空特征挖掘的交通流量预测方法

详细信息    查看全文 | 推荐本文 |

英文篇名：Traffic Flow Prediction Method Based on Spatio-Temporal Feature Mining 作者：孔繁钰 ; 周愉峰 ; 陈纲 英文作者：KONG Fan-yu;ZHOU Yu-feng;CHEN Gang;Chongqing Engineering Technology Research Center for Development Information Management,Chongqing Technology and Business University;Postdoctoral Research Station of Management Science and Engineering,Nanjing University of Aeronautics & Astronautics;College of Architecture and Urban Planning,Chongqing University; 关键词：深度神经网络 ; 改进卷积神经网络 ; 交通流量预测 ; 时空特征 ; 大数据 ; 自动学习 英文关键词：Deep neural network;;Improved convolutional neural network;;Traffic flow prediction;;Temporal-Spatial features;;Big data;;Automatic learning 中文刊名：JSJA 英文刊名：Computer Science 机构：重庆工商大学重庆市发展信息管理工程技术研究中心;南京航空航天大学管理科学与工程博士后流动站;重庆大学建筑城规学院; 出版日期：2019-07-15 出版单位：计算机科学 年：2019 期：v.46 基金：国家自然科学基金(71702015);; 中国博士后科学基金(2017M611810);; 重庆市社科规划重大应用项目(2017ZDYY51);; 重庆市发展信息管理工程技术研究中心开放基金项目(gczxkf201706);; 重庆工商大学科研平台开放课题(KFJJ2018078)资助 语种：中文; 页：JSJA201907049 页数：5 CN：07 ISSN：50-1075/TP 分类号：328-332

摘要

基于神经网络和大数据的交通流量预测方法层出不穷,但对交通流量预测的精度仍有待进一步提高。为了解决该问题,提出一种基于时空特征挖掘的交通流量预测方法。该方法使用改进的CNN来挖掘交通流量的空间特征,使用递归神经网络来挖掘交通流量的时间特征,能够充分利用交通流量的每周/每天的周期性和时空特征。此外,在该方法中还使用了一种基于相关性的模型,它可以根据过去的交通流量实现自动学习。实验结果表明,相比于其他几种较新的预测方法,所提方法具有较高的交通流量预测精度。
        Traffic forecasting methods using neural networks and big data are emerging in an endless stream,but their prediction accuracy for traffic flow is usually inaccurate.In order to solve this problem,this paper proposed a traffic flow forecasting method based on spatio-temporal feature mining.This method makes use of improving convolutional neural network(CNN) to mine the spatial features of traffic flow,and utilizes recursive neural network to mine the temporal features of traffic flow,so that it can make full use of weekly/daily periodicity and spatial-temporal characteristics of traffic flow.In addition,the method also introduces a correlation-based model that can achieve automatic learning according to the past traffic flow.Experiment results show that the proposed method has higher prediction accuracy for traffic flow compared with some novel methods.

引文

[1] RUI L L,LI Q M.Short-term Traffic Flow Prediction Algo- rithm Based on Combined Model [J].Journal of Electronics & Information Technology,2016,38(5):1227-1233.(in Chinese)芮兰兰,李钦铭.基于组合模型的短时交通流量预测算法[J].电子与信息学报,2016,38(5):1227-1233.
    [2] ZHENG X,CHEN W,WANG P,et al.Big Data for Social Transportation[J].IEEE Transactions on Intelligent Transportation Systems,2016,17(3):620-630.
    [3] 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.
    [4] LI L,LI Y,LI Z.Efficient missing data imputing for traffic flow by considering temporal and spatial dependence[J].Tramsportation Research Part C Emerging Technologies,2013,34(9):108-120.
    [5] YU H B,SHEN Q,FENG G C.Introduce Numerical Solution to Visualize Convolutional Neuron Networks Based on Numerical Solution [J].Computer Science,2017,44(S1):146-150.(in Chinese)俞海宝,沈琦,冯国灿.在反卷积网络中引入数值解可视化卷积神经网络[J].计算机科学,2017,44(S1):146-150.
    [6] LUO J,JIANG Y,LIU X,et al.Multi-scale convolutional-recursive neural networks for RGB-D object recognition [J].Application Research of Computers,2017,34(9):2834-2837.(in Chinese)骆健,蒋旻,刘星,等.多尺度卷积递归神经网络的RGB-D物体识别[J].计算机应用研究,2017,34(9):2834-2837.
    [7] JIANG X,ADELI H.Dynamic Wavelet Neural Network Model for Traffic Flow Forecasting[J].Journal of Transportation Engineering,2005,131(10):771-779.
    [8] 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,2015,54(3):187-197.
    [9] QIAN W,YANG H H,SUN Y J.Kalman filtering traffic flow prediction research based on phase space re-construction [J].Computer Engineering and Applications,2016,52(14):37-41.(in Chinese)钱伟,杨慧慧,孙玉娟.相空间重构的卡尔曼滤波交通流预测研究[J].计算机工程与应用,2016,52(14):37-41.
    [10] CAO C T,LIN X H,XU L H.Short-term Traffic Flow Prediction Algorithm Based on FCM and Optimized SVR with Social Spider Optimization Algorithm [J].Journal of China Academy of Electronics and Information Technology,2017,12(1):52-59.(in Chinese)曹成涛,林晓辉,许伦辉.联合FCM与群集蜘蛛优化SVR的短时交通流量预测[J].中国电子科学研究院学报,2017,12(1):52-59.
    [11] ZHANG Y,HAGHANI A.A gradient boosting method to improve travel time prediction[J].Transportation Research Part C,2015,58(2):308-324.
    [12] HUANG W,SONG G,HONG H,et al.Deep Architecture for Traffic Flow Prediction:Deep Belief Networks with Multitask Learning[J].IEEE Transactions on Intelligent Transportation Systems,2014,15(5):2191-2201.
    [13] YU D,LIU Y,YU X.A Data Grouping CNN Algorithm for Short-Term Traffic Flow Forecasting[J].Web Technologies and Applications,2016,9931:92-103.
    [14] DUAN Y,LV Y,WANG F Y.Travel time prediction with LSTM neural network[C]//IEEE International Conference on Intelligent Transportation Systems.IEEE,2016:1053-1058.
    [15] CHEN Z H,LAN Y Y,GUO J F,et al.Distributed Stochastic Gradient Descent with Discriminative Aggregating [J].Chinese Journal of Computers,2015,38(10):2054-2063.(in Chinese)陈振宏,兰艳艳,郭嘉丰,等.基于差异合并的分布式随机梯度下降算法[J].计算机学报,2015,38(10):2054-2063.
    [16] WEI S,WYNTER L.Rejoinder:real-time road traffic forecasting using regime-switching space-time models and adaptive lasso[M].John Wiley and Sons Ltd,2012:297-315.
    [17] HAO Y,BAI Y P,ZHANG X F,et al.Application of Convolution Neural Network in SAR Target Recognition[J].Journal of Chongqing University of Technology(Natural Science),2018,32(5):210-215.(in Chinese)郝岩,艳萍,张校非,等.卷积神经网络在SAR目标识别中的应用[J].重庆理工大学学报(自然科学),2018,32(5):210-215.