中文核心期刊
CSCD来源期刊
中国科技核心期刊
RCCSE中国核心学术期刊

Journal of Chongqing Jiaotong University(Natural Science) ›› 2020, Vol. 39 ›› Issue (11): 11-19.DOI: 10.3969/j.issn.1674-0696.2020.11.02

• Transport+Big Data and Artificial Intelligence • Previous Articles     Next Articles

Prediction Method of Bus Passenger Flow Based on RS-IPSOSVM

HUANG Yishao1,2, HAN Lei2,3   

  1. (1. Engineering Research Center of Catastrophic Prophylaxis and Treatment of Road & Traffic Safety, Ministry of Education, Changsha University of Science & Technology, Changsha 410114, Hunan, China; 2. School of Traffic and Transportation Engineering, Changsha University of Science & Technology, Changsha 410114, Hunan, China; 3. Key Laboratory of Road & Traffic Engineering of Ministry of Education, Tongji University, Shanghai 201804, China)
  • Received:2019-03-25 Revised:2019-08-19 Online:2020-11-19 Published:2020-11-23

基于RS-IPSOSVM的公交客流量预测方法

黄益绍1,2,韩磊2,3   

  1. (1. 长沙理工大学 道路灾变防治及交通安全教育部工程研究中心,湖南 长沙 410114; 2. 长沙理工大学 交通运输工程学院,湖南 长沙 410114; 3. 同济大学 道路与交通工程教育部实验室,上海 201804)
  • 作者简介:黄益绍(1976—),男,湖南郴州人,副教授,博士,主要从事交通控制、交通规划方面的研究。E-mail:744861302@qq.com 通信作者:韩磊(1994—),男,安徽六安人,博士研究生,主要从事深度学习、短时交通流预测、智能交通方面的研究。E-mail:meleihan@163.com
  • 基金资助:
    湖南省自然科学基金项目(2018JJ2444);湖南省教育厅科学研究重点项目(16A007);长沙理工大学道路灾变防治及交通安全教育部工程研究中心开放基金资助项目(kfj140401)

Abstract: In order to improve the accuracy of bus passenger flow forecasting, based on bus IC card data, the variation law and the influencing factors of bus passenger flow were explored, and a prediction method of bus passenger flow based on rough set (RS), improved particle swarm optimization (IPSO) and support vector machine (SVM) was proposed. Firstly, the influencing factors of bus passenger flow were determined by deep mining of passenger flow data. Secondly, rough set was used to reduce the attributes of 13 initial influencing factors and eliminate redundant information to get 8 core influencing factors.Thirdly, the adaptive inertia weight and asynchronous learning factors were introduced to optimize the PSO algorithm, and the IPSO algorithm was used to find the global optimal parameters of the SVM. The kernel function was used to map the core influencing factors of bus passenger flow into high-dimensional space, and the nonlinear mapping relationship between the core influence factors and the bus passenger flow was fitted to realize the passenger flow prediction. Finally, the proposed method was verified by the data of bus passenger flow in Guangzhou. The results show that the prediction accuracy of the proposed method is above 90%, which simplifies the training samples, overcomes the blindness of SVM parameter selection, and effectively improves the practicability and reliability.

Key words: transportation engineering, bus passenger flow prediction, rough sets, improved particle swarm optimization, support vector machine, data mining

摘要: 为了提高公交客流预测的准确性,基于公交IC卡数据,挖掘公交客流的变化规律和影响因素,提出了一种基于粗糙集(RS)和改进粒子群(IPSO)优化支持向量机(SVM)的公交客流量预测方法。首先,通过对客流数据的深度挖掘,确定公交客流的影响因子;其次,利用粗糙集对13个初始影响因子进行属性约简,剔除冗余信息,得到8个核心影响因子;再次,引入自适应调整的惯性权重和异步变化的学习因子对PSO算法进行优化,利用IPSO算法来寻找SVM全局最优参数,通过核函数将公交客流核心影响因子映射到高维空间,拟合核心影响因子与公交客流量间的非线性映射关系,实现客流的预测;最后,以广州市公交线路客流数据进行了方法验证。结果表明:所用方法预测精度在90%以上,简化了训练样本,克服了SVM参数选择的盲目性,实用性和可靠性均得到有效提高。

关键词: 交通工程, 公交客流量预测, 粗糙集, 改进粒子群优化, 支持向量机, 数据挖掘

CLC Number: