System and method for alternative destination recommendation on ridesharing platforms

1. A computer-implemented method for recommending alternative destinations, comprising:

receiving, by a computing device of a ridesharing platform from a terminal device, a trip request for a rider that comprises an origin and a destination;

classifying, by the computing device of the ridesharing platform, the trip request into one of a plurality of trip purpose categories based at least on the origin and the destination of the trip request, the rider's information, and a classifier trained to predict the one trip purpose category of the trip request;

in response to the one trip purpose category belonging to a preset group of trip purpose categories:

determining, by the computing device of the ridesharing platform based on the one trip purpose category, one or more alternative destinations for the trip request, wherein the one or more alternative destinations share one or more features with the destination of the trip request;

ranking the one or more alternative destinations; and

displaying, in addition to the destination from the trip request, the ranked one or more alternative destinations on the terminal device;

wherein the preset group of trip purpose categories is determined by at least:

identifying one or more historical trips in which alternative destinations were recommended and accepted by one or more historical users; and

adding one or more trip purpose categories corresponding to the one or more historical trips into the preset group of trip purpose categories.

2. The computer-implemented method of claim 1 , wherein the classifier is trained based on training data comprising a plurality of historical trips labeled with a plurality of trip purpose categories, wherein for each of the plurality of historical trips, the training data comprises one or more of the following: trip information of the historical trip, rider information of the historical trip, point-of-interest information of the historical trip, and a label representing a trip purpose category of the historical trip.

3. The method of claim 2 , wherein the training data are obtained at least partially by:

automatically sending, by the computing device of the ridesharing platform to a training rider's computing device, a query for the training rider to input a trip purpose category of a training trip that the training rider took; and

labeling the training trip with the trip purpose category inputted by the training rider.

4. The method of claim 2 , wherein the point-of-interest information of the historical trip comprises at least one of the following, for each of the plurality of historical trips comprising a destination:

a plurality of point-of-interests (POIs) within a preset range of the destination, wherein the plurality of POIs are ranked by popularity.

5. The method of claim 2 , wherein the trip information of the historical trip comprises a destination of the historical trip, and the rider information of the historical trip comprises at least one of the following:

whether the destination is a home location or a work location;

a travel frequency to the destination on weekdays; and

a travel frequency to the destination on weekends.

6. The method of claim 2 , wherein the trip information of the historical trip for training the classifier comprises at least one of the following:

day-of-week of the trip;

a start time of the trip; and

a travel time duration of the trip.

7. The method of claim 1 , wherein the classifier is trained as one of the following models: Random Forest (RF), Deep Neural Network (DNN), XGBoost, and logistic regression.

8. The method of claim 1 , wherein the determining one or more alternative destinations for the trip request comprises:

determining an estimated cost of the trip request, a service level of the trip request, and a business type of the destination of the trip request;

identifying a plurality of point-of-interest (POI) locations of the business type within a range of the origin of the trip request;

determining, for each of the POI locations, an estimated cost of a hypothetical trip with the service level from the origin to the POI location;

determining one or more of the POIs locations with corresponding estimated costs of the hypothetical trips that are not greater than the estimated cost of the trip request; and

identifying the one or more alternative destinations for the trip request from the one or more determined POI locations.

9. The method of claim 8 , wherein the service level comprises at least one of the following:

trip configuration of the trip request comprising solo trip or carpool trip; and

vehicle configuration of the trip request comprising vehicle capacity or vehicle class.

10. The method of claim 1 , wherein the ranking comprises:

ranking the alternative destinations based on surge multipliers or matching probabilities of the alternative destinations.

11. The method of claim 1 , wherein the receiving a trip request that comprises an origin and a destination comprises:

receiving the trip request comprising an initial origin from the terminal device;

determining one or more alternative origins based on the initial origin;

displaying the one or more alternative origin to the terminal device for the rider to select; and

determining one of the one or more alternative origins selected by the rider as the origin of the trip request.

12. A system comprising one or more processors and one or more non-transitory computer-readable memories coupled to the one or more processors, the one or more non-transitory computer-readable memories storing instructions that, when executed by the one or more processors, cause the system to perform operations comprising:

receiving, by a computing device of a ridesharing platform from a terminal device, a trip request for a rider that comprises an origin and a destination;

classifying, by the computing device of the ridesharing platform, the trip request into one of a plurality of trip purpose categories based at least on the origin and the destination of the trip request, the rider's information, and a classifier trained to predict the one trip purpose category of the trip;

in response to the one trip purpose category belonging to a preset group of trip purpose categories:

determining, by the computing device of the ridesharing platform based on the one trip purpose category, one or more alternative destinations for the trip request, wherein the one or more alternative destinations share one or more features with the destination of the trip request;

ranking the one or more alternative destinations; and

displaying, in addition to the destination from the trip request, the ranked one or more alternative destinations on the terminal device;

wherein the preset group of trip purpose categories is determined by at least:

identifying one or more historical trips in which alternative destinations were recommended and accepted by one or more historical users; and

adding one or more trip purpose categories corresponding to the one or more historical trips into the preset group of trip purpose categories.

13. The system of claim 12 , wherein the determining one or more alternative destinations for the trip request comprises:

determining an estimated cost of the trip request, a service level of the trip request, and a business type of the destination of the trip request;

identifying a plurality of point-of-interest (POI) locations of the business type within a range of the origin of the trip request;

determining, for each of the POI locations, an estimated cost of a hypothetical trip with the service level from the origin to the POI location;

determining one or more of the POI locations with corresponding estimated costs of the hypothetical trips that are not greater than the estimated cost of the trip request; and

identifying the one or more alternative destinations for the trip request from the one or more determined POI locations.

14. The system of claim 12 , the operations further comprise:

identifying one or more historical trips in which alternative destinations were selected by a user; and

adding one or more trip purpose categories corresponding to the one or more identified historical trips to the preset group of trip purpose categories.

15. The system of claim 12 , wherein the classifier is trained based on training data comprising a plurality of historical trips labeled with a plurality of trip purpose categories, wherein for each of the plurality of historical trips, the training data comprises one or more of the following: trip information of the historical trip, rider information of the historical trip, point-of-interest information of the historical trip, and a label representing a trip purpose category of the historical trip.

16. A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:

receiving, by a computing device of a ridesharing platform from a terminal device, a trip request for a rider that comprises an origin and a destination;

classifying, by the computing device of the ridesharing platform, the trip request into one of a plurality of trip purpose categories based at least on the origin and the destination of the trip request, the rider's information, and a classifier trained to predict the one trip purpose category of the trip;

in response to the one trip purpose category belonging to a preset group of trip purpose categories:

determining, by the computing device of the ridesharing platform based on the one trip purpose category, one or more alternative destinations for the trip request, wherein the one or more alternative destinations share one or more features with the destination of the trip request;

ranking the one or more alternative destinations; and

displaying, in addition to the destination from the trip request, the ranked one or more alternative destinations on the terminal device;

wherein the preset group of trip purpose categories is determined by at least:

identifying one or more historical trips in which alternative destinations were recommended and accepted by one or more historical users; and

adding one or more trip purpose categories corresponding to the one or more historical trips into the preset group of trip purpose categories.

17. The storage medium of claim 16 , wherein the determining one or more alternative destinations for the trip request comprises:

determining an estimated cost of the trip request, a service level of the trip request, and a business type of the destination of the trip request;

identifying a plurality of point-of-interest (POI) locations of the business type within a range of the origin of the trip request;

determining, for each of the POI locations, an estimated cost of a hypothetical trip with the service level from the origin to the POI location;

determining one or more of the POI locations with corresponding estimated costs of the hypothetical trips that are not greater than the estimated cost of the trip request; and

identifying the one or more alternative destinations for the trip request from the one or more determined POI locations.

18. The storage medium of claim 16 , the operations further comprise:

identifying one or more historical trips in which alternative destinations were selected by a user; and

adding one or more trip purpose categories corresponding to the one or more identified historical trips to the preset group of trip purpose categories.

19. The storage medium of claim 16 , wherein the classifier is trained based on training data comprising a plurality of historical trips labeled with a plurality of trip purpose categories, wherein for each of the plurality of historical trips, the training data comprises one or more of the following: trip information of the historical trip, rider information of the historical trip, point-of-interest information of the historical trip, and a label representing a trip purpose category of the historical trip.