Short-term FFBS demand prediction with multi-source data in a hybrid deep learning framework

被引:22
作者
Bao, Jie [1 ,2 ]
Yu, Hao [1 ,2 ]
Wu, Jiaming [1 ,2 ]
机构
[1] Southeast Univ, Jiangsu Key Lab Urban ITS, Nanjing, Jiangsu, Peoples R China
[2] Southeast Univ, Sch Transportat, Dongnandaxue 2, Nanjing, Jiangsu, Peoples R China
基金
中国国家自然科学基金; 美国国家科学基金会;
关键词
transportation; road traffic; bicycles; learning (artificial intelligence); autoregressive moving average processes; regression analysis; convolutional neural nets; recurrent neural nets; bike-sharing service provider; commuting pattern; bike lanes; prediction performance; short-term memory; convolutional neural network; artificial neural network; nonpeak hours; evening peak; morning peak; HDL-net exhibits better performance; hybrid deep learning neural network; weekend rides; weekday rides; spatiotemporal patterns; air quality data; hourly aggregated weather data; Shanghai city; FFBS trip data; deep learning approach; free-floating bike sharing; short-term travel demand; hybrid deep learning framework; multisource data; short-term FFBS demand prediction; NEURAL-NETWORK; BIKE;
D O I
10.1049/iet-its.2019.0008
中图分类号
TM [电工技术]; TN [电子技术、通信技术];
学科分类号
0808 ; 0809 ;
摘要
The primary objective of this study is to predict the short-term demand of free-floating bike sharing (FFBS) using deep learning approach. The FFBS trip data in Shanghai city are collected from the Mobike Company. Other datasets such as weather data and air quality data are also collected. The spatiotemporal patterns of FFBS demand indicates that the weekday rides exhibit an obvious commuting pattern while the weekend rides are usually involved with various trip purposes. Then, a hybrid deep learning neural network (HDL-net) is developed to predict the short-term FFBS demand for different time intervals including 15, 20 and 30 min. The proposed HDL-net exhibits better performance on morning peak than the evening peak and non-peak hours for all the three time intervals. Moreover, five benchmark methods are also used to compare with the proposed HDL-net. The results suggest that the proposed hybrid deep learning framework outperform the benchmarks in the prediction performance for all three time intervals. The results of this study could provide insightful suggestions for transportation authorities to develop effective rebalancing strategies and bike lanes planning schemes to promote the service level of cycling in an urban city.
引用
收藏
页码:1340 / 1347
页数:8
相关论文
共 33 条
[1]  
[Anonymous], P 97 TRB ANN M WASH
[2]  
[Anonymous], P 23 ACM SIGKDD INT
[3]  
[Anonymous], P 2017 IEEE INT C BI
[4]  
[Anonymous], SUSTAINABILITY BASEL
[5]  
[Anonymous], 2015, P 29 ANN C NEUR INF
[6]  
[Anonymous], 2017, MARK SHAR MOB IS APP
[7]  
[Anonymous], ACCIDENT ANAL PREV
[8]   Better Understanding of Factors Influencing Likelihood of Using Shared Bicycle Systems and Frequency of Use [J].
Bachand-Marleau, Julie ;
Lee, Brian H. Y. ;
El-Geneidy, Ahmed M. .
TRANSPORTATION RESEARCH RECORD, 2012, (2314) :66-71
[9]   Spatial Analysis of Bikeshare Ridership With Smart Card and POI Data Using Geographically Weighted Regression Method [J].
Bao, Jie ;
Shi, Xiaomeng ;
Zhang, Hao .
IEEE ACCESS, 2018, 6 :76049-76059
[10]   Exploring Bikesharing Travel Patterns and Trip Purposes Using Smart Card Data and Online Point of Interests [J].
Bao, Jie ;
Xu, Chengcheng ;
Liu, Pan ;
Wang, Wei .
NETWORKS & SPATIAL ECONOMICS, 2017, 17 (04) :1231-1253