Real-Time Return Demand Prediction Based on Multisource Data of One-Way Carsharing Systems
Table 1
Characteristic variables and symbols.
Category
Variables
Symbols
Examples
Station set
Whether the station () in the current trajectory point’s station set () belongs to the stations where the user has picked up vehicles
whetherO
1
Whether the station () in the current trajectory point’s station set () belongs to the stations where the user has returned vehicles
whetherD
1
Driving mileage
Current track length + 15 min mileage > user’s historical minimum mileage
whetherLumin
1
Current track length + 15 min mileage user’s historical maximum mileage
whetherLumax
1
User historical average mileage
Luave
27.26 km
User historical median mileage
Lumedian
25.53 km
Current track length+15 min mileage > minimum mileage of the user’s historical order with the same starting point
whetherLuomin
1
Current track length+15 min mileagemaximum mileage of the user’s historical order with the same starting point
whetherLuomax
1
The historical average mileage of user’s order with the same starting point
Luoave
27.56 km
Historical median mileage of user’s order with the same starting point
Luomedian
26.77 km
Current track length + 15 min mileage
LaddR
20.15 km
Driving duration
Current travel duration+15 min > user’s historical minimum travel duration
whetherTumin
1
Current travel duration+15minuser’s historical maximum travel duration
whetherTumax
1
Current travel duration+15 min> the 15th percentile of the user’s historical travel duration
whetherTuq15
1
Current travel duration+15minthe 85th percentile of the user’s historical travel duration
whetherTuq85
1
User historical average driving duration
Tuave
15.10 min
User historical median driving duration
Tumedian
14.70 min
Current travel duration+15 min > minimum travel duration of the user’s historical order with the same starting point
whetherTuomin
1
Current travel duration+15 min maximum travel duration of the user’s historical order with the same starting point
whetherTuomax
1
Current travel duration+15 min> the 15th percentile of travel duration of the user’s historical order with the same starting point
whetherTuoq15
1
Current travel duration+15 min the 85th percentile of travel duration of the user’s historical order with the same starting point
whetherTuoq85
1
User average driving duration of the user’s historical order with the same starting point
Tuoave
14.32 min
User median driving duration of the user’s historical order with the same starting point
Tuomedian
15.22 min
Current travel duration+15 min
Tadd15
16 min
EAPP
A user opened the APP within 15 minutes before returning the vehicle in history
APP
0
Destination station
The number of all users’ historical orders whose destination station can be matched with station () in the station set ()
Dordernum
1
The proportion of all users’ historical orders whose destination station can be matched with station () in the station set ()
Dorderfre
0.2
The minimum time difference between the order duration of all users’ historical orders whose destination station can be matched with station () in the station set () and the current trajectory duration ()
DdetminT
3.3 min
The number of the user’s historical orders whose destination station can be matched with station () in the station set ()
UDordernum
4
The proportion of the user’ historical orders whose destination station can be matched with station () in the station set ()
UDorderfre
0.25
The minimum time difference between the order duration of the user’s historical orders whose destination station can be matched with station () in the station set () and the current trajectory duration ()
UDdetminT
4.1 min
Pick-up station and destination station
The number of all users’ historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set ()
ODordernum
2
The proportion of all users’ historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set ()
ODorderfre
0.1
The minimum time difference between the order duration of all users’ historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set () and the current trajectory duration ()
ODdetminT
6.3 min
The number of the user’ historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set ()
UODordernum
2
The proportion of the user’ historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set ()
UODorderfre
0.05
The minimum time difference between the order duration of the user’s historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set () and the current trajectory duration ()
UODdetminT
5.5 min
The trajectory difference between the order duration of the user’s historical orders whose starting station is the same as the current trajectory’s starting point and destination station can be matched with a station () in the station set () and the current trajectory duration ()