Aiming at the problem of tidal supply-demand imbalance of shared bicycles, a built environment indicator system was constructed from the three dimensions of socio-demographic, transportation design, and land use. Based on the Agglomerative Hierarchical Clustering model, typical time-aggregation patterns of shared bicycles were obtained from clustering. The nonlinear relationship and the threshold effect between them were revealed by using the Random Forest and SHapley Additive exPlanations models. Ultimately, Kunming was an example to prove that. The results show that the temporal aggregation modes of shared bicycles mainly show the first-in-last-out mode and the first-out-last-in mode, accounting for 91.4% of the sum of all modes. There is a significant nonlinear relationship between the built environment and the temporal aggregation modes of shared bicycles. The temporal aggregation modes of shared bicycles are primarily influenced by the density of the bus line network, proximity of parks and attractions, proximity of metro stations, and POI density of companies. When the proximity of metro stations is less than 0.5 km, there is an elevating effect on the probability of showing the first-in-last-out mode. When the POI density of companies is less than 25 per cell, there is a depressing effect on the probability of showing the first-in-last-out mode.
目前,建成环境对共享单车时空出行特征的影响逐步成为研究热点。现有研究主要集中于建成环境与共享单车出行量或到达量的关系,如Faghih等[3]考察了加拿大蒙特利尔市共享单车使用量的影响因素,发现离中央商务区越近,共享单车的出发量越大。黄颙昊等[4]以北京市路段作为研究单元,发现娱乐设施数量与共享单车流量之间存在负相关关系。马新卫等[5]以南京市为案例,发现娱乐型兴趣点(Point of interest,POI)在郊区和主城区对共享单车需求的影响存在异质性。上述研究多采用线性研究方法,多元线性回归模型被广泛用于研究建成环境对共享单车的使用影响重要程度[6];广义线性回归模型常被用于预测不同建成环境区域共享单车出行量的大小[7];空间自回归模型则被用于研究建成环境对共享单车出行的空间异质性影响[8]。然而,Ding等[9]的研究表明非线性效应将使建成环境的作用机理发生变化;高楹等[10]发现建成环境与共享单车集聚间存在非线性关系,而现有研究普遍忽略了建成环境对共享单车使用的非线性作用。同时,现有研究局限于分别研究建成环境与共享单车出行量或到达量的关系,难以精确刻画土地利用对共享单车时间集聚模式的影响,无法为解决共享单车潮汐式供需失衡问题提供有效的理论支撑。
相似度度量方法层面,由于欧氏距离易受到时序曲线振幅的影响,难以反映其形状和趋势动态变化的相似性,因此并不适用;同时,由于时间序列的长度相同,而动态时间规整(Dynamic time warping, DTW)方法是一种非对称的相似度度量方法,其只考虑时间序列之间的对齐,而不考虑它们之间的顺序关系,早晚高峰变化趋势相反的时间序列易归为一类,因此也不适用。本文采用布雷柯蒂斯相异度(Bray-curtis dissimilarity,BCD)来衡量样本之间的距离[15]。布雷柯蒂斯相异度是一种常用于衡量时间序列相似性的指标,其具有能衡量时间序列如何随时间共同变化、可解释性强、计算高效等优势,能够较好地反映共享单车净流入量时间序列的动态变化趋势相似性。布雷柯蒂斯相异度的取值在[0,1]之间,取值越接近1表示两时间序列间的相异程度越高。布雷柯蒂斯相异度计算公式为:
KuangJia-heng, WuQun-yong. Spatial-temporal equilibrium analysis and attraction area optimization of dockless sharing bicycles connected to subway stations[J]. Journal of Geo-information Science, 2022, 24(7):1337-1348.
[4]
Faghih ImaniA, EluruN, El-GeneidyA M, et al. How land-use and urban form impact bicycle flows: Evidence from the bicycle-sharing system(BIXI) in montreal[J]. Journal of Transport Geography, 2014, 41: 306-314.
HuangYong-hao, YangXin-miao, YueJin-tao. Urban street bicycle flow analysis based on multi-scale geographically weighted regression model[J]. Journal of Tsinghua University(Science and Technology), 2022, 62(7): 1132-1141.
MaXin-wei, JiYan-jie, JinYu-chuan, et al. Geographically and temporally weighted regression for modeling spatio-temporal variation in dockless bikeshare usage demand[J]. Journal of Jilin University (Engineering and Technology Edition), 2020, 50(4):1344-1354.
[9]
JiY, MaX, YangM, et al. Exploring spatially varying influences on metro-bikeshare transfer: A geographically weighted poisson regression approach[J]. Sustainability, 2018, 10(5): No.1526.
[10]
WangK, ChenY J. Joint Analysis of the impacts of built environment on bikeshare station capacity and trip attractions[J]. Journal of Transport Geography, 2020, 82: No.102603.
[11]
ShenY, ZhangX, ZhaoJ. Understanding the usage of dockless bike sharing in Singapore[J]. International Journal of Sustainable Transportation, 2018, 12(9): 686-700.
[12]
DingC, ChenP, JiaoJ. Non-linear effects of the built environment on automobile-involved pedestrian crash frequency: a machine learning approach[J]. Accident Analysis & Prevention, 2018, 112: 116-126.
GaoYing, SongCi, GuoSi-hui, et al. Spatial-temporal characteristics and influencing factors of source and sink of dockless sharing bicycles connected to subway stations[J]. Journal of Geo-information Science, 2021, 23(1): 155-170.
[15]
CaoM, CaiB, MaS, et al. Analysis of the cycling flow between origin and destination for dockless shared bicycles based on singular value decomposition[J]. ISPRS International Journal of Geo-Information, 2019, 8(12): 573.
LiWu, JIXiao-feng, ChenFang, et al. Nonlinear influence of built environment on spatial fairness of conventional transit[J]. Journal of Highway and Transportation Research and Development, 2023, 40(8): 207-213.
[18]
EwingR, CerveroR. Travel and the built environment: A synthesis[J]. Transportation Research Record, 2001, 1780(1): 87-114.
WangJian-jun, WangSai, SongMing-yang, et al. Review of research on impact of urban built environment on travel behavior in big data context[J]. Journal of Chang'an University(Natural Science Edition),2022, 42(1): 61-78.
[21]
RodionovI V, SozontovA N. 2020. On confidence estimation based on quantitative similarity coefficients[J]. Automation and Remote Control, 81(2): 320-332.
LiDong-mei, YangYu, MengXiang-hao, et al. Review on multi-lable classification[J]. Journal of Frontiers of Computer Science and Technology, 2023,17(11):2529-2542.
[24]
AasK, JullumM, LolandA. Explaining individual predictions when features are dependent: More accurate approximations to Shapley values[J]. Artificial Intelligence, 2021, 298: No.103502.