Trip reranking for a journey planner

1. A method for re-ranking trips on an associated transportation network, comprising:

receiving, from a plurality of automatic ticketing validation systems, validation information for each of a plurality of travelers of the associated transportation network, the validation information including a timestamp, at least one location, and a ticket identification;

extracting, from the validation information, a trip for each of the plurality of travelers of the transportation network, the trip comprising an origin, a destination, and at least one of a departure time or a service;

storing the extracted trips in an associated database;

receiving, at a computer system from a user device over a computer network, a trip request from an associated user, the trip request including an origin, a destination and a departure time on the associated transportation network;

retrieving, from the associated database, a set of actual trips corresponding to the received trip request, the set of actual trips each having a common origin, destination and departure time corresponding to the received trip request;

determining a ranking function associated with the trip request in accordance with the retrieved set of actual trips corresponding to the received trip request;

receiving, from an associated journey planner, a list of candidate trips on the associated transportation network corresponding to the received trip request;

applying the ranking function to the list of candidate trips so as to re-rank the candidate trips in the list thereof;

returning, to the user device over the computer network, a re-ranked list of candidate trips in response to the trip request; and

displaying, on a display of the user device, the re-ranked list of candidate trips,

wherein at least one of the receiving, retrieving, determining, applying, returning is performed by a computer processor of the computer system.

2. The method of claim 1 , wherein the retrieving the set of actual trips further comprises developing, for each trip in the set thereof, a set of time-variant features characterizing the trip on the associated transportation network.

3. The method of claim 2 , further comprising converting the set of actual trips into a set of real-world traveler preferences in accordance with the set of time-variant features, the set of real-world traveler preferences corresponding to implicit preferences of travelers of the associated transportation network.

4. The method of claim 3 , wherein the time-variant features comprise at least one of a global feature, a local feature, or a dynamic feature.

5. The method of claim 3 , wherein determining the ranking function further comprises, for each trip in the set of actual trips matching the received trip request:

receiving, from the journey planner, a set of trip candidates corresponding to the trip request;

analyzing the set of top trip candidates to determine whether the trip of the set of actual trips has been preferred to any of candidate trips;

adding the trip of the set of actual trips to a set of training trips associated with the trip request responsive to a preference of the trip over the candidate trips; and

learning a ranking function corresponding to the trip request in accordance with the set of training trips.

6. The method of claim 5 , wherein the ranking function is a ranking support vector machine ranking function.

7. The method of claim 1 , wherein the associated transportation network is a public transportation network.

8. The method of claim 1 , wherein the set of actual trips corresponding to the received trip request is generated from the extracted trips in the associated database.

9. A system comprising memory storing instructions for performing the method of claim 1 , and a processor in communication with the memory which implements the instructions.

10. A system for re-ranking trips on an associated transportation network, comprising:

a journey planner including a journey planning engine configured to receive a trip request and return a list of candidate trips on the associated transportation network responsive thereto;

a re-ranking component configured to apply a ranking function to the list of trips and generate a re-ranked list of candidate trips;

a processor of a computer system in communication with a user device over a computer network; and

memory, in communication with the processor, which stores instructions which are executed by the processor for:

receiving, from a plurality of automatic ticketing validation systems, validation information for each of a plurality of travelers of the associated transportation network, the validation information including a timestamp, at least one location, and a ticket identification,

extracting, from the validation information, a trip for each of the plurality of travelers of the transportation network, the trip comprising an origin, a destination, and at least one of a departure time or a service,

storing the extracted trips in an associated database,

retrieving, from the associated database, a set of actual trips corresponding to the received trip request, the set of actual trips each having a common origin, destination and departure time corresponding to the received trip request,

for each trip in the set of actual trips:

receiving, from the journey planner, a set of trip candidates corresponding to the trip request,

analyzing the set of top trip candidates to determine whether the trip of the set of actual trips has been preferred to any of candidate trips,

adding the trip of the set of actual trips to a set of training trips associated with the trip request responsive to a preference of the trip over the candidate trips,

learning a ranking function corresponding to the trip request in accordance with the set of training trips,

returning, to the user device over the computer network, the re-ranked list of candidate trips in accordance with the learned ranking function, and

displaying, on the user device the re-ranked list of candidate trips in accordance with the learned ranking function.

11. The system of claim 10 , further comprising instructions for developing, for each trip in the set of actual trips, a set of time-variant features characterizing the trip on the associated transportation network.

12. The system of claim 11 , further comprising instructions for converting the set of actual trips into a set of real-world traveler preferences in accordance with the set of time-variant features, the set of real-world traveler preferences corresponding to implicit preferences of travelers of the associated transportation network.

13. The system of claim 12 , wherein the time-variant features comprise at least one of a global feature, a local feature, or a dynamic feature.

14. The system of claim 10 , wherein the ranking function is a support vector machine function.

15. The system of claim 12 , wherein the set of actual trips corresponding to the received trip request is generated from the extracted trips in the associated database.

16. A computer-implemented method for re-ranking trips on an associated public transportation network, comprising:

receiving, from a plurality of automatic ticketing validation systems, validation information for each of a plurality of travelers of the transportation network, the validation information including a timestamp, at least one location, and a ticket identification;

extracting, from the validation information, a trip for each of the plurality of travelers of the transportation network, the trip comprising an origin, a destination, and at least one of a departure time or a service;

storing the extracted trips in an associated database;

receiving a trip request from an associated user device over a computer network, the trip request including an origin, a destination and a departure time on the associated transportation network;

retrieving, from the associated database, a set of actual trips corresponding to the received trip request, the set of actual trips each having a common origin, destination and departure time corresponding to the received trip request, wherein the set of actual trips corresponding to the received trip request is generated from the extracted trips in the associated database;

for each trip in the set of actual trips matching the received trip request:

receiving, from the journey planner, a set of trip candidates corresponding to the trip request, analyzing the set of top trip candidates to determine whether the trip of the set of actual trips has been preferred to any of candidate trips,

adding the trip of the set of actual trips to a set of training trips associated with the trip request responsive to a preference of the trip over the candidate trips,

developing, for each trip in the set thereof, a set of time-variant features characterizing the trip on the associated transportation network,

converting the set of actual trips into a set of real-world traveler preferences in accordance with the set of time-variant features, the set of real-world traveler preferences corresponding to implicit preferences of travelers of the associated transportation network, and

learning a ranking function corresponding to the trip request in accordance with the set of training trips and real-world traveler preferences;

receiving, from an associated journey planner, a list of candidate trips on the associated transportation network corresponding to the received trip request;

applying the ranking function to the list of candidate trips so as to re-rank the candidate trips in the list thereof;

returning, in response to the trip request, a re-ranked list of candidate trips to the user device over the computer network; and

displaying, on the user device the re-ranked list of candidate trips.