Application US20190066130

UNIFIED ARTIFICIAL INTELLIGENCE MODEL FOR MULTIPLE CUSTOMER VALUE VARIABLE PREDICTION
[+] Show Details

US20190066130

1. A method for creating a unified model for predicting customer value using artificial intelligence, comprising: (8) (4)

creating, at a computer system, a series of two or more neural network modules, each of the neural network modules including a dense layer of neurons connected to all of the neurons for the immediately preceding neural network module;

creating, at the computer system, a plurality of variable sub-task neural network modules, wherein each of the variable sub-task neural network modules is connected to an output of the last of the series of two or more neural network modules, and wherein each of the variable sub-task neural network modules is configured to calculate a separate one of a plurality of component variables for predicted customer value; and

creating, at the computer system, a final task neural network module for calculating predicted customer value, wherein the final task neural network module is connected to the output of the last of the series of two or more neural network modules and is also connected to an intermediate output from each of the plurality of variable sub-task neural network modules;

wherein the unified model is configured to predict, using particular input data pertaining to a particular customer, each of the plurality of component variables for predicted customer value and also separately predict total customer value for that customer.

2. The method of claim 1, wherein the unified model further comprises an input layer connected to one or more of the two or more neural network modules. (0)

3. The method of claim 1, wherein the two or more neural network modules each includes a normalization component and a dropout component. (0)

4. The method of claim 1, wherein each of the plurality of variable sub-task neural network modules comprises: (0) (3)

5. The method of claim 1, further comprising: (1) (1)

7. The method of claim 1, wherein the plurality of component variables for predicted customer value consist of: (0) (4)

8. The method of claim 1, further comprising: (0) (2)

9. The method of claim 1, wherein the plurality of component variables for predicted customer value comprises loss for a given customer, including predicted losses from fraud. (0)

10. The method of claim 1, wherein the plurality of component variables for predicted customer value comprises cost for a given customer, including predicted costs charged to an electronic payment transaction service provider by financial institutions from which money is received and/or to which money is sent. (0)

11. A non-transitory computer-readable medium having stored thereon instructions that are executable by a computer system to cause the computer system to perform operations comprising: (5) (4)

creating a series of two or more neural network modules, each of the neural network modules including a dense layer of neurons connected to all of the neurons for the immediately preceding neural network module;

creating a plurality of variable sub-task neural network modules, wherein each of the variable sub-task neural network modules is connected to an output of the last of the series of two or more neural network modules, and wherein each of the variable sub-task neural network modules is configured to calculate a separate one of a plurality of component variables for predicted customer value; and

creating a final task neural network module for calculating predicted customer value, wherein the final task neural network module is connected to the output of the last of the series of two or more neural network modules and is also connected to an intermediate output from each of the plurality of variable sub-task neural network modules;

12. The non-transitory computer-readable medium of claim 11, wherein the unified model further comprises an input layer connected to one or more of the two or more neural network modules. (0)

13. The non-transitory computer-readable medium of claim 11, wherein the two or more neural network modules each includes a normalization component and a dropout component. (0)

14. The non-transitory computer-readable medium of claim 11, wherein each of the plurality of variable sub-task neural network modules comprises: (0) (3)

15. The non-transitory computer-readable medium of claim 11, wherein the plurality of component variables for predicted customer value consist of: (0) (4)

16. The non-transitory computer-readable medium of claim 11, wherein the operations further comprise: (0) (2)

17. A computer system, comprising: (3) (2)

a processor; and

one or more computer-readable media having stored thereon instructions that are executable to cause the computer system to perform operations comprising: creating a series of two or more neural network modules, each of the neural network modules including a dense layer of neurons connected to all of the neurons for the immediately preceding neural network module;

18. The computer system of claim 17, wherein the unified model further comprises an input layer connected to one or more of the two or more neural network modules. (0)

19. The computer system of claim 17, wherein the two or more neural network modules each includes a normalization component and a dropout component. (0)

20. The computer system of claim 17, wherein each of the plurality of variable sub-task neural network modules comprises: (0) (3)

View Abstract and Specification Size

PDF with Images and Document Face >

Full Text Publication >

Application US20190066130

UNIFIED ARTIFICIAL INTELLIGENCE MODEL FOR MULTIPLE CUSTOMER VALUE VARIABLE PREDICTION

Patent Matrix® Search