Please submit manuscripts in either of the following two submission systems

    ScholarOne Manuscripts

  • ScholarOne
  • 勤云稿件系统

  • 登录

Search by Issue

  • 2024 Vol.31
  • 2023 Vol.30
  • 2022 Vol.29
  • 2021 Vol.28
  • 2020 Vol.27
  • 2019 Vol.26
  • 2018 Vol.25
  • 2017 Vol.24
  • 2016 vol.23
  • 2015 vol.22
  • 2014 vol.21
  • 2013 vol.20
  • 2012 vol.19
  • 2011 vol.18
  • 2010 vol.17
  • 2009 vol.16
  • No.1
  • No.2

Supervised by Ministry of Industry and Information Technology of The People's Republic of China Sponsored by Harbin Institute of Technology Editor-in-chief Yu Zhou ISSNISSN 1005-9113 CNCN 23-1378/T

期刊网站二维码
微信公众号二维码
Related citation:Wen Zheng,Wenquan Li,Qian Chen,Yan Zheng,Chenhao Zhang.Trip Purposes of Automobile Users Inference Using Multi-day Traffic Monitoring Data[J].Journal of Harbin Institute Of Technology(New Series),2023,30(5):1-11.DOI:10.11916/j.issn.1005-9113.2022099.
【Print】   【HTML】   【PDF download】   View/Add Comment  Download reader   Close
←Previous|Next→ Back Issue    Advanced Search
This paper has been: browsed 2280times   downloaded 1620times 本文二维码信息
码上扫一扫!
Shared by: Wechat More
Trip Purposes of Automobile Users Inference Using Multi-day Traffic Monitoring Data
Author NameAffiliation
Wen Zheng School of Transportation, Southeast University, Nanjing 211189, China 
Wenquan Li School of Transportation, Southeast University, Nanjing 211189, China 
Qian Chen School of Transportation, Southeast University, Nanjing 211189, China 
Yan Zheng School of Transportation, Southeast University, Nanjing 211189, China 
Chenhao Zhang School of Transportation, Southeast University, Nanjing 211189, China 
Abstract:
Determining trip purpose is an important link to explore travel rules. In this paper, we take automobile users in urban areas as the research object, combine unsupervised learning and supervised learning methods to analyze their travel characteristics, and focus on the classification and prediction of automobile users’ trip purposes. However, previous studies on trip purposes mainly focused on questionnaires and GPS data, which cannot well reflect the characteristics of automobile travel. In order to avoid the multi-day behavior variability and unobservable heterogeneity of individual characteristics ignored in traditional traffic questionnaires, traffic monitoring data from the Northern District of Qingdao are used, and the K-means clustering method is applied to estimate the trip purposes of automobile users. Then, Adaptive Boosting (AdaBoost) and Random Forest (RF) methods are used to classify and predict trip purposes. Finally, the result shows: (1) the purpose of automobile users can be mainly divided into four clusters, which include Commuting trips, Flexible life demand travel in daytime,Evening entertainment and leisure shopping, and Taxi-based trips for the first three types of purposes, respectively; (2) the Random Forest method performs significantly better than AdaBoost in trip purpose prediction for higher accuracy; (3) the average prediction accuracy of Random Forest under hyper-parameters optimization reaches 96.25%, which proves the feasibility and rationality of the above clustering results.
Key words:  trip purpose, automobile users, traffic monitoring data, K-means clustering, AdaBoost, random forest
DOI:10.11916/j.issn.1005-9113.2022099
Clc Number:U491.1 + 22
Fund:
Descriptions in Chinese:
  

基于多日交通监测数据推断汽车用户的出行目的

郑文,李文权,陈茜,郑炎,张晨皓

东南大学交通学院,211189,中国南京

摘要:确定出行目的是探索出行规律的一个重要环节。本文以城市中的汽车用户为研究对象,结合无监督学习和有监督学习方法分析其出行特征,重点研究汽车用户出行目的的分类和预测。然而,以往关于出行目的的研究主要集中在问卷调查和GPS数据上,不能很好地反映汽车出行的特点。为了避免传统交通问卷调查中忽略的多日行为变异性和个体特征的不可观察异质性,我们采用了青岛市北区的道路卡口监测数据,并运用K-均值聚类方法来估计汽车用户的出行目的。然后,采用自适应提升(AdaBoost)和随机森林(RF)方法对出行目的进行分类和预测。最后,本研究的结果显示: (1)汽车用户的出行目的主要可以分为四类,分别是通勤出行、出租车出行、白天的弹性生活需求出行和晚间的娱乐休闲购物;(2)随机森林方法在出行目的预测中的表现明显优于AdaBoost,准确率更高;(3)超参数优化下随机森林的平均预测精度达到96.25%,证明了上述聚类结果的可行性和合理性。

关键词:出行目的;汽车用户;交通监控数据;K-means聚类; AdaBoost;随机森林

主要研究内容:

由于汽车出行结构的复杂性,使用GPS数据或问卷调查数据无法完全分析出其出行特征。本文选择覆盖面更广的道路卡口监测数据来分析汽车出行的空间和时间特征。本文以K-means聚类和机器学习为重点,旨在对汽车用户的出行目的进行分类和预测。我们的工作可以为探索城市汽车出行的特征提供依据,也可以为不同的出行目的提供合适的公共交通方式(如需求响应式公交、快速公交和公交快线)来替代汽车出行,这有助于优化城市道路交通结构,保证城市交通系统的可持续发展。其核心思想和重要研究内容主要包括以下三个部分。

1)在我们的研究中,应用了以往关于出行目的研究中没有使用过的道路卡口监测数据,并验证了监督和非监督学习方法的结合在分析具体出行问题上的有效性。这不仅可以从一个新的角度来分析出行目的,也为今后的出行行为研究提供了很好的参考。

2)基于青岛市的道路监测数据集,建立了六个指标来分析汽车用户出行的时空特征。然后,采用K-means聚类方法对数据集进行分析,根据SSE和CH_Score的收敛值,可以将研究区域内的汽车用户出行目的主要分为四类,分别为通勤出行、出租车出行、白天的弹性生活需求出行和晚间的娱乐休闲购物。这使得我们可以间接了解汽车用户的出行结构,从而帮助交通管理者针对不同目的实施不同的交通管理措施。

3)根据K-means聚类的结果,从多个方面分析了两种基于树的集合学习方法的预测精度。结果发现,在超参数优化条件下,随机森林更适合于处理该类出行目的预测问题。MAE、RMSE和R2的值都优于AdaBoost。此外,RF的平均预测精度可以达到96.25%,也高于AdaBoost。它为处理出行行为问题提供了参考。

LINKS