厦门市大型公交站点共享单车接驳需求预测模型分析  

Connections Demand Forecast Model for Shared Bicycle Around Large-scale Public Transport Stations in Xiamen

作　　者：林子旸 LIN Ziyang(Xiamen Land Space and Transport Research Center,Xiamen 361001,China)

机构地区：[1]厦门市国土空间和交通研究中心,厦门361001

出　　处：《交通与运输》2021年第4期15-19,共5页Traffic & Transportation

摘　　要：近年来共享单车逐渐成为大型公共交通方式接驳的重要手段,但站点周边需求不一、投放过量的问题也成为管理部门的困扰。目前的需求预测方法存在基础数据不全、无法验证泛化能力等问题,难以支撑精细化的需求预测。基于厦门市长时间、全样本的共享单车数据,结合详实的人口、建成环境、天气等基础数据,分别对大型公交站点为起点和终点的共享单车出行需求建立传统数学模型、机器学习模型、深度学习模型,研究大型公交站点的共享单车接驳需求的关键影响指标。并以厦门市轨道2号线开通后站点周边流量数据为例,进行模型校核,通过情景假设,对环境优化后的骑行情况作量化分析。In recent years,shared bicycle has gradually become an important way for people to connect between large public transportation station,but the problem of different needs around the site and excessive investment has also become problems for management departments.Current demand forecasting methods have problems such as incomplete basic data and inability to verify generalization capabilities,and it is difficult to support refined demand forecasting.Based on the long-term,full-sample bicycle-sharing data of Xiamen,combined with detailed population,built environment,weather and other basic data,this paper establishes traditional algorithms,machine learning,and deep learning models for the bicycle-sharing traveling demand.At the same time,using scenario hypothesis on Xiamen to validate models and get quantitative analysis of optimizing riding environment situations.

关 键 词：城市交通 共享单车 轨道接驳 机器学习模型 大型公交站点 

分 类 号：U491[交通运输工程—交通运输规划与管理]