共享单车停放需求影响因素分析和预测研究

为了解决城市共享单车的乱停乱放问题,本文基于北京市的共享单车出行大数据,提出了共享单车停放需求预测的多项Logit模型。首先分析了单车停放需求的影响因素,然后选取了时间、空间及天气方面的12个因素为自变量,通过Wald检验分析了这些因素与停放需求的相关性和显著性,基于多项Logit模型建立了共享单车的停放需求预测模型。结果表明:工作日、时段、商业区、所临道路类型、临近轨交站、高温、下雨、以及风力等级与共享单车停放需求显著相关;构建的预测模型总体预测准确率为77.5%,其中对出现频率最高的低停放需求预测准确率高达86.49%。

Analysis of Influencing Factors and Prediction for the Parking Demand of Bike Sharing

In order to solve the problem of disorderly parking of Free-Floating Bike Sharing(FFBS) in cities,a Multinomial Logit Model to forecast the parking demand of FFBS is proposed in this paper based on the bicycle trip data in Beijing.Firstly,the influencing factors of FFBS parking demand were analyzed.Then,12 factors in three aspects:time,space and weather were selected as independent variables.Wald test was used to analyze the correlation and significance of these factors to parking demand.Finally,based on Multinomial Logit Model,a parking demand prediction model for FFBS was established.The results show that:working day,time period,commercial,type of adjacent road,adjacent to transit hub,high temperature,rain,and wind scale,these 8 factors are significantly correlated with the parking demand of FFBS.The overall prediction accuracy of the prediction model is 77.5%,especially the prediction accuracy of low parking demand is the highest,reaching 86.49%.