Granted Patent B2

US 12,346,744 · App. 17/414,317 · Granted Jul 1, 2025

Inventor: Kyosuke Tomoda (Tokyo, JP)

Assignee: RAKUTEN GROUP, INC.

G06F9/5077G06F9/5016

Abstract

Claims (51)

1. A prediction system, comprising at least one processor configured to:

acquire a plurality of candidates of information influencing obtaining a desired result in a predetermined service to a user;

predict, for each of the plurality of candidates, based on a machine learning model trained on parameters which include a set of candidates information, whether the desired result is obtained;

identify the candidates predicted to obtain the desired result

wherein the at least one processor is configured to predict whether the reference is satisfied based on at least one of a setting of a first system used in the predetermined service, a situation predicted for the first system, a setting of a second system on which the first system depends, and a situation predicted for the second system, and based on the candidate;

wherein the at least one processor is configured to provide the identified candidate predicted to satisfy the reference;

wherein the identified candidate is a resource candidate comprising at least one hardware resource;

wherein the predetermined service utilizes the identified candidate to obtain the desired result;

wherein the desired result being greater than or equal to a threshold value for a minimum response speed; and

wherein the desired result is that the user continues to use the predetermined service.

2. The prediction system according to claim 1 ,

wherein as a value indicated by the information becomes higher, the desired result is more likely obtained and a cost relating to the predetermined service becomes higher, and

wherein the at least one processor is configured to identify the candidate which is predicted to obtain the desired result, and which has a lower cost than another candidate cost.

3. The prediction system according to claim 2 , wherein the at least one processor is configured to identify the candidate having a lowest cost among the candidates predicted to obtain the desired result.

4. The prediction system according to claim 1 , wherein the at least one processor is configured to predict, for each of the plurality of candidates, whether the desired result is obtained by using a prediction model trained based on training data in which the information is input and whether the desired result is obtained is output.

5. The prediction system of claim 1 , where the least one processor is further configured to identify the candidates not predicted to obtain the desired result.

6. The prediction system of claim 1 , wherein the desired result is also greater than or equal to a threshold value for a communication speed.

7. A prediction system, comprising at least one processor configured to:

acquire a plurality of candidates of information influencing obtaining a desired result in a predetermined service;

predict, for each of the plurality of candidates, based on a machine learning model trained on parameters which include a set of candidates information, whether the desired result is obtained; and

identify the candidates predicted to obtain the desired result,

wherein the desired result is a greater than or equal to a threshold value for a minimum response speed,

wherein the information indicates a privilege to be conferred to a user of the predetermined service,

wherein the desired result is that the user continues to use the predetermined service,

wherein the at least one processor is configured to predict, for each of the plurality of candidates, whether the user continues to use the predetermined service,

wherein the at least one processor is configured to identify the candidate for which it is predicted that the user continues to use the predetermined service;

wherein the identified candidate is a resource candidate comprising at least one hardware resource; and

wherein the predetermined service utilizes the identified candidate to obtain the desired result.

8. The prediction system according to claim 7 ,

wherein the at least one processor is configured to calculate a defection rate of the user from the predetermined service based on a predetermined calculation method;

wherein the at least one processor is configured to calculate a maximum value of the privilege conferrable to the user based on the defection rate,

wherein the at least one processor is configured to identify, within a range of the maximum value, the candidate for which it is predicted that the user continues to use the predetermined service.

9. The prediction system according to claim 8 , wherein the at least one processor is configured to determine the defection rate based on at least one of an attribute of the user, a usage situation of the predetermined service by the user, and a usage situation of another service by the user.

10. The prediction system according to claim 7 ,

wherein the at least one processor is configured to calculate an expected profit of the user based on a predetermined calculation method, and

wherein the at least one processor is configured to identify, based on the expected profit, the candidate for which it is predicted that the user continues to use the predetermined service.

11. The prediction system according to claim 10 ,

wherein the at least one processor is configured to calculate a maximum value of the privilege conferrable to the user based on the expected profit, and

wherein the at least one processor is configured to identify, within a range of the maximum value corresponding to the expected profit, the candidate for which it is predicted that the user continues to use the predetermined service.

12. The prediction system according to claim 7 , wherein the at least one processor is configured to predict whether the user continues to use the predetermined service based on at least one of a defection rate of the user from the predetermined service, an attribute of the user, a usage situation of the predetermined service by the user, and a usage situation of another service by the user.

13. A non-transitory computer-readable information storage medium for storing a program for causing a computer to:

acquire a plurality of candidates of information influencing obtaining a desired result in a predetermined service;

predict, for each of the plurality of candidates, based on a machine learning model trained on parameters which include a set of candidates information, whether the desired result is obtained; and

identify the candidates predicted to obtain the desired result,

wherein the desired result is a greater than or equal to a threshold value for a minimum response speed,

wherein the information indicates a privilege to be conferred to a user of the predetermined service,

wherein the desired result is that the user continues to use the predetermined service,

wherein the program causes the computer to predict, for each of the plurality of candidates, whether the user continues to use the predetermined service,

wherein the program causes the computer to identify the candidate for which it is predicted that the user continues to use the predetermined service;

wherein the identified candidate is a resource candidate comprising at least one hardware resource; and

wherein the predetermined service utilizes the identified candidate to obtain the desired result.

Assignments (2)

Continuity (1)

References Cited (24)