多种缺失模式下交通数据组合近似填补方法

doi:10.16097/j.cnki.1009-6744.2023.06.005

交通运输系统工程与信息 ›› 2023, Vol. 23 ›› Issue (6): 42-50.DOI: 10.16097/j.cnki.1009-6744.2023.06.005

• 智能交通系统与信息技术 • 上一篇下一篇

多种缺失模式下交通数据组合近似填补方法

郭凤香¹，黄金涛¹，陈昱光¹，郭延永²，刘攀^*2

1. 昆明理工大学，交通工程学院，昆明 650504；2. 东南大学，交通学院，南京 210096

收稿日期:2023-08-09 修回日期:2023-08-28 接受日期:2023-09-12 出版日期:2023-12-25 发布日期:2023-12-23
作者简介:郭凤香(1979- )，女，黑龙江海林人，教授，博士
基金资助:
国家杰出青年科学基金(51925801)；国家重点研发计划 (2018YFB1600900)

Combined Approximate Method for Traffic Data Imputation Under Multiple Missing Modes

GUO Feng-xiang¹,HUANG Jin-tao¹,CHEN Yu-guang¹,GUO Yan-yong²,LIU Pan^*2

1. Faculty of Transportation Engineering, Kunming University of Science and Technology, Kunming 650504, China; 2. School of Transportation, Southeast University, Nanjing 210096, China

Received:2023-08-09 Revised:2023-08-28 Accepted:2023-09-12 Online:2023-12-25 Published:2023-12-23
Supported by:
National Science Fund for Distinguished Young Scholars of China (51925801)；National Key Research and Development Program of China (2018YFB1600900)

摘要/Abstract

摘要： 随着智能交通系统中采集和应用的基础数据规模不断扩大，数据缺失问题的重要性也日益凸显。针对交通数据中常出现的数据随机缺失和连续缺失问题，本文提出基于鲸鱼优化算法优化最小二乘支持向量机的组合近似填补方法(Combined Approximate Filling, CAF)。考虑缺失数据整体变化趋势的同时，参考数据的波动特征，根据多重填补思想对缺失值分别使用单变量填补和多变量填补，然后引入图片识别中自适应阈值分割法对不同时段下的差异值进行动态划分处理，最后利用不同时段的动态差异度阈值将单变量填补和多变量填补的结果进行结合，完成缺失值的高精度近似填补。为验证填补方法的性能，利用云南省玉溪市大量实车轨迹处理数据设计多组实验。实验结果表明，在小样本数据中，CAF填补方法能够适应多种场合的填补工作，该方法总体优于其他方法，在不同缺失率下均表现良好，尤其是随机缺失填补，最大 RMSE 为0.365。实验还证明了该方法在不同缺失类型和不同数据离散度下数据填补效果相比于其他方法优势更加明显。

关键词: 智能交通, 数据填补, 最小二乘支持向量机, 轨迹数据, 差异度

Abstract: With the continuous expansion of the scale of basic data collected and applied in intelligent transportation systems, the importance of missing data imputation has become increasingly prominent. Aiming at the problem of random missing and continuous missing data in traffic dataset, this paper proposes a Combined Approximate Filling (CAF) method based on least squares support vector machine optimized by whale optimization algorithm. Considering the overall trend of the missing data and the fluctuation characteristics of the reference data, this paper uses the univariate imputation and multivariate imputation to fill the missing values based on the multiple imputation theory. The adaptive threshold segmentation method in image recognition is introduced to dynamically divide the difference values under different periods. The results of univariate imputation and multivariate imputation are combined through the dynamic dissimilarity threshold under different periods to complete the high- precision approximate imputation. Several experiments are designed to verify the performance of the imputation method, using a large number of real vehicle trajectory data in Yuxi City of Yunnan Province in China. The results show that the CAF imputation method can adapt to various occasions in small sample data. This method produces satisfied results under different missing rates, especially for random missing imputation. The maximum RMSE is 0.365. The experimental results also prove that the data imputation from the proposed method is more effective than some traditional methods under different missing types and different data dispersion conditions.

Key words: intelligent transportation;data imputation;least squares support vector machine;trajectory data, degree of difference

中图分类号:

U491

郭凤香, 黄金涛, 陈昱光, 郭延永, 刘攀. 多种缺失模式下交通数据组合近似填补方法[J]. 交通运输系统工程与信息, 2023, 23(6): 42-50.

GUO Feng-xiang, HUANG Jin-tao, CHEN Yu-guang, GUO Yan-yong, LIU Pan. Combined Approximate Method for Traffic Data Imputation Under Multiple Missing Modes[J]. Journal of Transportation Systems Engineering and Information Technology, 2023, 23(6): 42-50.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2023.06.005

http://www.tseit.org.cn/CN/Y2023/V23/I6/42

参考文献

[1] LI Y, LI Z, LI L. Missing traffic data: Comparison of imputation methods[J]. IET Intelligent Transport Systems, 2014, 8(1): 51-57.

[2] ZHANG S. Nearest neighbor selection for iteratively KNN imputation[J]. Journal of Systems and Software, 2012, 85(11): 2541-2552.

[3] CHENG S, LU F, PENG P, et al. Short-term traffic forecasting: An adaptive ST-KNN model that considers spatial heterogeneity[J]. Computers, Environment and Urban Systems, 2018, 71(1): 186-198.

[4] YU Q, JIBIN L, JIANG L. An improved ARIMA-based traffic anomaly detection algorithm for wireless sensor networks[J]. International Journal of Distributed Sensor Networks, 2016, 12(1): 1-9.

[5] FENG X, LING X, ZHENG H, et al. Adaptive multikernel SVM with spatial-temporal correlation for short-term traffic flow prediction[J]. IEEE Transactions on Intelligent Transportation Systems, 2018, 20(6): 2001- 2013.

[6] HABTIE A B, ABRAHAM A, MIDEKSO D. Artificial neural network based real-time urban road traffic state estimation framework[J]. Computational Intelligence in Wireless Sensor Networks: Recent Advances and Future Challenges, 2017, 676(1): 73-97.

[7] ZHANG Y, WEI X, ZHANG X, et al. Self-attention graph convolution residual network for traffic data completion [J]. IEEE Transactions on Big Data, 2022, 9(2): 528-541.

[8] YANG B, KANG Y, YUAN Y Y, et al. ST-LBAGAN: Spatio-temporal learnable bidirectional attention generative adversarial networks for missing traffic data imputation[J]. Knowledge-Based Systems, 2021, 215(10): 1-10.

[9] WU X, XU M, FANG J, et al. A multi-attention tensor completion network for spatiotemporal traffic data imputation[J]. IEEE Internet of Things Journal, 2022, 9 (20): 20203-20213.

[10] LEI M, LABBE A, WU Y, et al. Bayesian kernelized matrix factorization for spatiotemporal traffic data imputation and kriging[J]. IEEE Transactions on Intelligent Transportation Systems, 2022, 23(10): 18962- 18974.

[11] WU P L, DING M, ZHENG Y B. Spatiotemporal traffic data imputation by synergizing low tensor ring rank and nonlocal subspace regularization[J]. IET Intelligent Transport Systems, 2023, 17(9): 1908-1923.

[12] HUANG L, LI Z, LUO R, et al. Missing traffic data imputation with a linear generative model based on probabilistic principal component analysis[J]. Sensors, 2022, 23(1): 1-13.

[13] 成卫, 黄金涛, 陈昱光, 等. 基于浮动车速度波动特征的交通状态识别[J]. 交通运输系统工程与信息, 2023, 23(1): 67-76. [CHENG W, HUANG J T, CHEN Y G, et al. Traffic state recognition based on speed fluctuation characteristics of floating car[J]. Journal of Transportation Systems Engineering and Information Technology, 2023, 23(1): 67-76.]

[14] VAN BUUREN S. Flexible imputation of missing data (Second Edition)[M]. Boca Raton: CRC Press, 2018.

[15] ZHENG L, HUANG H, ZHU C, et al. A tensor-based K-nearest neighbors method for traffic speed prediction under data missing[J]. Transportmetrica B: Transport Dynamics, 2020, 8(1): 182-199.

多种缺失模式下交通数据组合近似填补方法

Combined Approximate Method for Traffic Data Imputation Under Multiple Missing Modes

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

[1]	陈峥, 赵文龙, 郭凤香, 赵志刚, 刘昱, 刘永刚. 计及安全性与舒适性的智能车辆换道轨迹规划研究[J]. 交通运输系统工程与信息, 2024, 24(1): 55-65.
[2]	裴玉龙, 傅博涵, 王子奇, 张杰. 引力理论框架下基于综合竞争力的自动驾驶拟人换道决策模型[J]. 交通运输系统工程与信息, 2024, 24(1): 66-80.
[3]	李传耀, 张帆, 王涛, 黄德鑫, 唐铁桥. 基于深度强化学习的道路交叉口生态驾驶策略研究[J]. 交通运输系统工程与信息, 2024, 24(1): 81-92.
[4]	王丹, 林业. 警告刺激对驾驶员接管性能的影响机理研究[J]. 交通运输系统工程与信息, 2024, 24(1): 106-114.
[5]	李岩, 王泰州, 徐金华, 陈姜会, 汪帆. 面向动态交通分配的交通需求深度学习预测方法[J]. 交通运输系统工程与信息, 2024, 24(1): 115-123.
[6]	任其亮, 徐韬, 刘媛, 程龙春. 考虑载客状态的改进孤立森林浮动车异常数据检测算法[J]. 交通运输系统工程与信息, 2024, 24(1): 124-131.
[7]	周备, 张莹, 张生瑞, 周千喜, 汪琴. 基于生成对抗网络的追尾事故数据填补方法研究[J]. 交通运输系统工程与信息, 2024, 24(1): 132-137.
[8]	袁见, 刘福强, 安琨, 郑喆, 马万经, 俞秋田. 基于交通波特征的车道级车流溯源方法[J]. 交通运输系统工程与信息, 2024, 24(1): 159-167.
[9]	胡立伟, 张瑞杰, 赵雪亭, 贺雨, 陈琛, 刘冰, 侯智. 道路环形交叉口机动车冲突风险区识别模型研究[J]. 交通运输系统工程与信息, 2024, 24(1): 168-178.
[10]	曲大义, 孟奕名, 王韬, 宋慧, 陈意成. 基于分子力场的网联自主车辆跟驰安全特性及模型[J]. 交通运输系统工程与信息, 2023, 23(6): 33-41.
[11]	蒋贤才, 徐慧智. 混合驾驶环境下交叉口空间资源动态控制方法[J]. 交通运输系统工程与信息, 2023, 23(6): 63-73.
[12]	李泰国, 张天策, 李超, 周星宏. 基于面部倒立摆模型与信息熵的驾驶员疲劳检测[J]. 交通运输系统工程与信息, 2023, 23(5): 24-32.
[13]	蒋贤才, 程国柱. 网联交通环境下交叉口进口车道动态配置方法[J]. 交通运输系统工程与信息, 2023, 23(5): 55-66.
[14]	李鹏程, 董宝田, 李思贤. 基于时间图注意力的交叉口交通状态识别及关联度研究[J]. 交通运输系统工程与信息, 2023, 23(5): 67-74.
[15]	蔡浩, 李林峰, 李涵, 李新, 周腾. 基于极限学习机的短期交通流预测混合优化模型[J]. 交通运输系统工程与信息, 2023, 23(5): 75-82.