摘要
通过对杭州地铁2019年1月1日到2019年1月25日的地铁刷卡数据进行分析,根据进出站高峰小时系数和站点位置将80个站点分为居住、工作、交通场站和混合类型四类。不同类型的车站早高峰晚高峰进出站高峰小时系数均不相同。对不同地铁线路的换乘量进行分析发现3号线换乘量比例最高,占其出站人数的77.7%。使用机器学习方法(随机森林和lightgbm)对不同站点每小时的进出站人数进行预测,平均相对误差均值为9.0%。表现出较强的可预测性。
Based on subway card data of Hangzhou metro on January 1,2019 to January 25,2019,according to the inbound and outbound of the station peak hour coefficient and site location,80 stations can be divided into living,working,traffic hub and the mixed type four categories.Different types of station morning peak and evening peak inbound and outbound of the station peak hour coefficient are not the same.By analyzing the transfer volume of different subway lines,it is found that the transfer volume of line 3 is the highest,accounting for 77.7%of the number of people leaving the station.The machine learning method(Randomforest and Lightgbm)is used to predict the number of people entering and leaving the station at different stations per hour.The average mean absolute percentage error(MAPE)is 9.0%,showing strong predictability.
作者
张素洁
谢小园
ZHANG Su-jie;XIE Xiao-yuan(Wuhan Railway Vocational College of Technology,Wuhan 430012,China)
出处
《价值工程》
2019年第19期65-67,共3页
Value Engineering
