System and method for generating predictions based on wireless commerce transactions

• US 10,068,251 B1
• Filed: 06/26/2008
• Issued: 09/04/2018
• Est. Priority Date: 06/26/2008
• Status: Active Grant

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Scofield, Christopher L.
• Primary Examiner(s): Brown, Luis A
• Application Number:
  US12/147,162
• Time in Patent Office: 3,722 Days
• Field of Search: 705 1449, 705 1453, 705 1464, 705 1466, 705 1467, 725 34, 725 62, 4554141, 455466, 4554121
• CPC Class Codes: G06Q 30/0251 Targeted advertisements
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute

Abstract

Various embodiments of a system and method for generating predictions based on wireless commerce transactions are described. Various embodiments may include a wireless commerce transaction analysis system configured to generate one or more models based on aggregated wireless commerce transaction information indicating one or more wireless commerce transactions completed with respective mobile devices of multiple users. For each of at least some of the wireless commerce transactions, the aggregated wireless commerce transaction information may indicate one or more characteristics of the wireless commerce transaction. The wireless commerce transaction analysis system may also be configured to generate a prediction of a future action to be performed by a user of a particular mobile device. The generation of the prediction may be dependent upon an evaluation of the one or more models and one or more wireless commerce transactions completed with the particular mobile device at one or more locations.

Claims

1. A system, comprising:

a network interface configured to receive wireless transaction data from one or more transaction reporters;

a processor; and

a memory coupled to the processor, wherein the memory comprises program instructions executable by the processor to;

generate one or more models based on aggregated wireless transaction information indicating sessions that include respective sequences of multiple wireless transactions completed with respective mobile communication devices of a plurality of users, wherein for individual ones of the multiple wireless transactions, an identification number for a user of a mobile communication device and authentication information of the user is transmitted from the mobile communication device to a transaction terminal used for the wireless transaction when the mobile communication device is within a given proximity of the transaction terminal, and wherein the aggregated wireless transaction information is based at least on the wireless transaction data, wherein

at least one wireless transaction of at least one of the sequences of wireless transactions is completed using a corresponding mobile device by a user of the mobile device, and wherein

the generation of the one or more models comprises;

determination of a plurality of paths from individual ones of said wireless transactions to respective other ones of said wireless transactions, wherein a given path from a transaction to another transaction is determined based on a given mobile device completing the other transaction within a given period of time after completing the transaction, and wherein individual paths of said plurality of paths have respective path weights, wherein a given path that has been established a greater number of times than another given path has a higher path weight than the other given path; and

identification of optimal aggregate paths within the plurality of paths, wherein individual ones of the optimal aggregate paths comprises a series of two or more paths between different wireless transactions,

wherein a given aggregate path is identified as an optimal aggregate path out of a plurality of aggregate paths between two wireless transactions by determining that a sum of individual path weights of the given aggregate path is higher than a sum of individual path weights of other ones of the aggregate paths, and wherein the other ones of the aggregate paths are identified as non-optimal aggregate paths,

and wherein the one or more models comprises a representation of a graph of the plurality of paths, wherein the graph includes multiple levels, each level comprising one or more of said wireless transactions, and wherein to determine a given optimal aggregate path out of two or more aggregate paths to a particular wireless transaction of a given level, one or more aggregate paths previously determined to be non-optimal aggregate paths to one or more respective wireless transactions of a previous level are disregarded;

generate a prediction of a future action to be performed by a user of a particular mobile communication device, wherein

generation of the prediction is dependent upon an evaluation of the one or more generated models and a current session including a path representing a sequence of multiple wireless transactions completed with the particular mobile communication device, and

the generation of the prediction comprises selection of the future action based at least in part on;

the identification of the optimal aggregate paths between different ones of said wireless transactions, and

a determination of the wireless transactions that have been completed with the particular mobile communication device;

determine, based on said prediction of the future action, information to transmit to the particular mobile communication device; and

transmit, via a network, said determined information to the particular mobile communication device.

2. The system of claim 1, wherein the information communicated to the particular mobile communication device comprises an advertisement based on the generated prediction.

3. The system of claim 1, wherein to generate a prediction of a future action to be performed the program instructions are executable to generate a prediction as to an amount of money to be spent on an item by said user of said particular mobile device.

4. The system of claim 1, wherein said aggregated wireless transaction information indicates one or more wireless transactions completed via radio-frequency identification (RFID).

5. The system of claim 1, wherein to generate the one or more models based on aggregated wireless transaction information, the program instructions are executable to create said aggregated wireless transaction information from wireless transaction information reported by one or more transaction terminals.

6. The system of claim 1, wherein to generate the one or more models based on aggregated wireless transaction information the program instructions are executable to create said aggregated wireless transaction information from wireless transaction information reported by one or more of said respective mobile communication devices of the plurality of users.

7. The system of claim 1, wherein for individual ones of at least some of said wireless transactions, the generation of the one or more models comprises determination of one or more characteristics of the individual wireless transaction, wherein the one or more characteristics of the individual wireless transaction comprise one or more of:

a user identifier identifying a user of the mobile communication device used to complete said individual wireless transaction, a device identifier identifying the mobile communication device used to complete said individual wireless transaction, payment information associated with said individual wireless transaction, an item identifier identifying an item exchanged in said individual wireless transaction, a timestamp indicating a time at which said individual wireless transaction is completed, an identifier identifying the location of at which the individual wireless transaction is completed, information indicating the physical location at which the individual wireless transaction is completed, or information indicating the direction of travel of the user of the particular mobile communication device.

8. The system of claim 7, wherein said item identifier comprises a stock keeping unit (SKU).

9. The system of claim 1, wherein to generate a prediction of a future action to be performed, the program instructions are executable to generate a prediction as to an item to be purchased by said user of said particular mobile device.

10. The system of claim 1, wherein to generate a prediction of a future action to be performed the program instructions are executable to generate a prediction as to a location of to be visited by said user of said particular mobile device.

11. A computer-implemented method, comprising:

performing, by one or more computers;

generating one or more models based on aggregated transaction information indicating sessions that include respective sequences of multiple transactions completed by a plurality of users, wherein

one or more of said transactions is a wireless transaction completed using a corresponding mobile device by a user of the mobile device, and wherein

the generation of the one or models comprises;

a determination of a plurality of paths from the wireless transaction to respective other ones of the wireless transactions, wherein a given path from the transaction to another transaction is determined based on a given mobile device completing the other transaction within a given period of time after completing the transaction, and wherein individual paths of said plurality of paths have respective path weights, wherein a given path that has been established a greater number of times than another given path has a higher path weight than the other given path; and

identification of optimal aggregate paths within the plurality of paths, wherein individual ones of the optimal aggregate paths comprises a series of two or more paths between different wireless transactions,

wherein a given aggregate path is identified as an optimal aggregate path out of a plurality of aggregate paths between two wireless transactions by determining that a sum of individual path weights of the given aggregate path is higher than a sum of individual path weights of other ones of the aggregate paths, and wherein the other ones of the aggregate paths are identified as non-optimal aggregate paths,

and wherein the one or more models comprises a representation of a graph of the plurality of paths, wherein the graph includes multiple levels, each level comprising at least the wireless transaction or one or more of the other wireless transactions, and wherein to determine a given optimal aggregate path of two or more aggregate paths to a particular wireless transaction of a given level, one or more aggregate paths previously determined to be non-optimal aggregate paths to one or more respective wireless transactions of a previous level are disregarded;

identifying a mobile communication device based on transaction information of one or more other transactions completed by a particular user of the mobile communication device;

generating, by a processor of the one or more computers, a prediction of a future action to be performed by the particular user of the identified mobile communication device, wherein

generation of the prediction is dependent upon an evaluation of the one or more generated models and a current session including a path representing a sequence of multiple transactions completed by the particular user, and wherein

the generation of the prediction comprises selection of the future action based at least in part on;

the identification of the optimal aggregate paths between the wireless transaction and different ones of the other wireless transactions, and

a determination of the transactions that have been completed by the particular user of the identified mobile communication device;

determining, based on said prediction of the future action, information to transmit to the particular mobile communication device; and

transmitting, via a network, said determined information to the identified mobile communication device.

12. The method of claim 11, wherein the information communicated to the identified mobile communication device comprises an advertisement based on the generated prediction.

13. The method of claim 11, wherein at least a portion of said aggregated transaction information comprises wireless transaction information.

14. The method of claim 13, wherein said wireless transaction information indicates one or more wireless transactions completed via radio-frequency identification (RFID).

15. The method of claim 11, wherein generating the one or more models based on aggregated transaction information comprises creating said aggregated transaction information from transaction information reported by one or more transaction terminals.

16. The method of claim 11, wherein for individual ones of at least some of said transactions, the aggregated transaction information indicates one or more characteristics of the individual transaction, wherein the one or more characteristics of the individual transaction comprise one or more of:

a user identifier identifying a user completed said individual transaction, a device identifier identifying a mobile communication device associated with said user, payment information associated with said individual transaction, an item identifier identifying an item exchanged in said individual transaction, a timestamp indicating a time at which said individual transaction is completed, an identifier identifying the location of at which the individual transaction is completed, or information indicating the physical location at which the individual transaction is completed.

17. The method of claim 16, wherein said item identifier comprises a stock keeping unit (SKU).

18. The method of claim 11, wherein generating a prediction of a future action to be performed comprises generating a prediction as to an item to be purchased by said particular.

19. The method of claim 11, wherein generating a prediction of a future action to be performed comprises generating a prediction as to a location of to be visited by said particular user.

20. A non-transitory computer-readable storage medium, comprising program instructions computer-executable to:

generate one or more models based on aggregated wireless transaction information indicating sessions that include respective sequences of multiple wireless transactions completed with respective mobile communication devices of a plurality of users, wherein

at least one wireless transaction of at least one of said sequences of wireless transactions is completed using a corresponding mobile device by a user of the mobile device, and wherein

the generation of the one or models comprises;

a determination of a plurality of paths from the wireless transaction to a plurality of respective other ones of the wireless transactions, wherein a given path from the transaction to another transaction is determined based on a given mobile device completing the other transaction within a given period of time after completing the transaction, and wherein individual paths of said plurality of paths have respective path weights, wherein a given path that has been established a greater number of times than another given path has a higher path weight than the other given path; and

identification of optimal aggregate paths within the plurality of paths, wherein individual ones of the optimal aggregate paths comprises a series of two or more paths between different wireless transactions,

wherein a given aggregate path is identified as an optimal aggregate path out of a plurality of aggregate paths between two wireless transactions by determining that a sum of individual path weights of the given aggregate path is higher than a sum of individual path weights of other ones of the aggregate paths, and wherein the other ones of the aggregate paths are identified as non-optimal aggregate paths,

and wherein the one or more models comprises a representation of a graph of the plurality of paths, wherein the graph includes multiple levels, each level comprising at least the wireless transaction or one or more of the other wireless transactions, and wherein to determine a given optimal aggregate path to a particular wireless transaction of a given level, one or more aggregate paths previously determined to be non-optimal aggregate paths to one or more respective wireless transactions of a previous level are disregarded;

generate a prediction of a future action to be performed by a user of a particular mobile communication device, wherein

generation of the prediction is dependent upon an evaluation of the one or more generated models and a current session including a path representing a sequence of multiple wireless transactions completed with the particular mobile communication device, and wherein

the generation of the prediction comprises selection of the future action based at least in part on;

the identification of the optimal aggregate paths between the wireless transaction and different ones of the other wireless transactions, and

a determination of the transactions that have been completed with the particular mobile communication device;

determine, based on said prediction of the future action, information to transmit to the particular mobile communication device; and

transmit, via a network, said determined information to the particular mobile communication device.

21. The non-transitory computer-readable storage medium of claim 20, wherein the information communicated to the particular mobile communication device comprises an advertisement based on the generated prediction.

22. The non-transitory computer-readable storage medium of claim 20, wherein to generate a prediction of a future action to be performed the program instructions are executable to generate a prediction as to a location to be visited by said user of said particular mobile device.

23. The non-transitory computer-readable storage medium of claim 20, wherein said aggregated wireless transaction information indicates one or more wireless transactions completed via radio-frequency identification (RFID).

24. The non-transitory computer-readable storage medium of claim 20, wherein to generate the one or more models based on aggregated wireless transaction information, the program instructions are executable to create said aggregated wireless transaction information from wireless transaction information reported by one or more transaction terminals.

25. The non-transitory computer-readable storage medium of claim 20, wherein to generate the one or more models based on aggregated wireless transaction information the program instructions are executable to create said aggregated wireless transaction information from wireless transaction information reported by one or more of said respective mobile communication devices of the plurality of users.

26. The non-transitory computer-readable storage medium of claim 20, wherein for individual ones of at least some of said wireless transactions, the generation of the one or more models comprises determination of one or more characteristics of the individual wireless transaction, the one or more characteristics of the individual wireless transaction comprise one or more of:

a user identifier identifying a user of the mobile communication device used to complete said individual wireless transaction, a device identifier identifying the mobile communication device used to complete said individual wireless transaction, the payment information associated with said individual wireless transaction, an item identifier identifying an item exchanged in said individual wireless transaction, a timestamp indicating a time at which said individual wireless transaction is completed, an identifier identifying the location of at which the individual wireless transaction is completed, information indicating the physical location at which the individual wireless transaction is completed, or information indicating the direction of travel of the user of the particular mobile communication device.

27. The non-transitory computer-readable storage medium of claim 26, wherein said item identifier comprises a stock keeping unit (SKU).

28. The non-transitory computer-readable storage medium of claim 20, wherein to generate a prediction of a future action to be performed, the program instructions are executable to generate a prediction as to an item to be purchased by said user of said particular mobile device.

29. The non-transitory computer-readable storage medium of claim 20, wherein the generation of the one or models comprises a determination of weights indicating respective levels of user interest associated with respective ones of the paths.