基于因果分析图的城市交通流短时预测研究  

Predicting Short-Term Urban Traffics Based on Causality Analysis Graph

作　　者：王洁 高原[2] 张蕾[1] 马力文 冯筠[1] Wang Jie;Gao Yuan;Zhang Lei;Ma Liwen;Feng Jun(School of Information Science and Technology,Northwest University,Xi’an 710127,China;School of Economics and Management,Northwest University,Xi’an 710127,China;Network and Data Center,Northwest University,Xi’an 710127,China)

机构地区：[1]西北大学信息科学与技术学院,西安710127 [2]西北大学经济管理学院,西安710127 [3]西北大学网络和数据中心,西安710127

出　　处：《数据分析与知识发现》2022年第11期111-125,共15页Data Analysis and Knowledge Discovery

基　　金：国家社会科学基金面上项目(项目编号:20BTJ047)的研究成果之一。

摘　　要：【目的】有效地挖掘区域之间复杂的空间作用关系机制,提高短时交通流预测精度。【方法】提出一种新的图神经网络模型,该模型融合区域功能相似性矩阵与因果关系矩阵,按照“交通时序因果关系挖掘→时空特征提取→未来状态预测”的逻辑进行预测建模,训练图神经网络捕获区域内流量的时空依赖性特征,从而实现交通流量预测。【结果】在成都市滴滴出行数据集上进行实验分析,结果表明所提模型较其他8种基线模型效果均有一定的提升,相较于最优基线模型,在RMSE及MAE值上分别降低了3.098%和4.783%。【结论】交通时序因果图可以同时融合传统方法中通常需要考虑的空间距离关系特征、道路连通性特征、功能相似性特征等,且因果关系的引入能在一定程度上提升区域交通流的预测性能。[Objective]This paper examines the complex spatial interaction mechanism between regions,aiming to effectively predict short-term traffic flow.[Methods]Based on the graph neural network,we proposed a new predictive model,which integrated the regional functional similarity matrix and the causality matrix.Then,we developed a training strategy of“Mining traffic time series causal relationship→Extracting Spatio-temporal features→Predicting traffic flows”.Third,we predicted the traffic flows by capturing the spatio-temporal dependence characteristics of regional traffic.[Results]We tested the proposed model with Didi Chuxing data set from Chengdu.Compared with the optimal baseline model,the RMSE and MAE values were reduced by 3.098%and 4.783%,respectively.[Conclusions]The causal diagram for traffic sequence can simultaneously integrate the features of spatial distance relationships,road connectivity,and function similarities.With the help of causal relationships,the proposed model could effectively predict regional traffic flows.

关 键 词：短时交通流预测 交通时序因果图 时空数据 图神经网络 

分 类 号：U491[交通运输工程—交通运输规划与管理]