Systems and Methods to Identify Users Using an Automated Learning Process

• US 20110191768A1
• Filed: 02/03/2011
• Published: 08/04/2011
• Est. Priority Date: 02/03/2010
• Status: Active Grant

• Current Assignee: Yahoo Assets LLC
• Original Assignee: Xobni Corp. (Apollo Global Management, Inc.)
• Inventors: Smith, Adam Michael
• Granted Patent: US 8,924,956 B2
• Time in Patent Office: Days
• US Class Current: 717/176
• CPC Class Codes: G06F 8/61 Installation
• CPC Class Codes: G06F 8/65 Updates security arrangements therefor G06F21/57
• CPC Class Codes: G06F 9/445 Program loading or initiating bootstrapping G06F9/4401; security arrangements for program loading or initiating G06F21/57
• CPC Class Codes: G06N 20/00 Machine learning

Abstract

A computer-implemented method includes: collecting first data from first computers on which first software is installed, the first data including first characteristics associated with the first computers and adoption results of the first software; correlating the first characteristics with the adoption results to generate a correlation result; prior to installation of the first software on a second computer, collecting second data associated with characteristics of the second computer; and making a determination whether to install the first software on the second computer based on at least the second data and the correlation result.

Claims

1. A method, comprising:

collecting first data from first computers on which first software is installed, the first data including first characteristics associated with the first computers and adoption results of the first software on the first computers;

correlating, via a data processing system, the first characteristics with the adoption results to generate a correlation result;

prior to installation of the first software on a second computer, collecting second data associated with characteristics of the second computer; and

making a determination whether to install the first software on the second computer based on at least the second data and the correlation result.

2. The method of claim 1, wherein the correlation result is a model, and the correlating comprises providing the first characteristics and the adoption results as inputs to a machine learning algorithm to train the model, the first characteristics being parameters of the model and the adoption results being targets of the model.

3. The method of claim 1, wherein the correlating is performed using supervised machine learning.

4. The method of claim 1, further comprising:

correlating second characteristics associated with additional computers with adoption results of the first software on the additional computers to provide an update of the correlation result;

prior to installation of the first software on a third computer, collecting third data associated with characteristics of the third computer; and

making another determination whether to install the first software on the third computer based on at least the third data and the update of the correlation result.

5. The method of claim 1, wherein the first software is selected from the group consisting of software that implements a profile builder, software that implements a profile presenter, and software that implements both a profile builder and a profile presenter.

6. The method of claim 1, wherein the collecting the second data is performed via an installer during installation of a second software on the second computer, the method further comprising in response to the determination, offering the first software as a bundle with the second software.

7. The method of claim 6, wherein the second software is selected from the group consisting of a web browser, a document viewer, and a media player.

8. The method of claim 1, wherein the adoption results include a paid upgrade by a user from a first version to a second version of the first software.

9. The method of claim 1, wherein the adoption results include use of the first software by a user for more than a predetermined time period.

10. The method of claim 1, wherein the first characteristics are selected from the group consisting of hardware attributes of the first computers, software attributes of the first computers, user preferences for users of the first computers, usage patterns for the first computers, and combinations of the foregoing.

11. The method of claim 1, wherein the making the determination comprises predicting a likelihood of a particular adoption result, and determining whether the likelihood is above a threshold.

12. The method of claim 11, wherein the threshold is a dynamic threshold based on at least a cost of distribution of the first software.

13. The method of claim 1, wherein the making the determination comprises combining a likelihood of a plurality of adoption results to generate a value indicator to indicate a value for offering the first software.

14. A non-transitory computer-readable storage medium for tangibly storing thereon computer-readable instructions, the instructions causing a data processing system to perform a method, the method comprising:

collecting first data from first computers on which first software is installed, the first data including first characteristics associated with the first computers and adoption results of the first software on the first computers;

correlating, via the data processing system, the first characteristics with the adoption results to generate a correlation result;

prior to installation of the first software on a second computer, collecting second data associated with characteristics of the second computer; and

making a determination whether to install the first software on the second computer based on at least the second data and the correlation result.

15. The storage medium of claim 14, wherein the correlation result is a model, and the correlating comprises providing the first characteristics and the adoption results as inputs to a machine learning algorithm to train the model, the first characteristics being parameters of the model and the adoption results being targets of the model.

16. The storage medium of claim 14, wherein the first characteristics are selected from the group consisting of hardware attributes of the first computers, software attributes of the first computers, user preferences for users of the first computers, usage patterns for the first computers, and combinations of the foregoing.

17. The storage medium of claim 14, wherein the making the determination comprises predicting a likelihood of a particular adoption result, and determining whether the likelihood is above a threshold.

18. A system, comprising:

memory to store first data from first computers on which first software is installed, the first data including first characteristics associated with the first computers and adoption results of the first software on the first computers, and to further store second data associated with characteristics of a second computer; and

at least one processor coupled to access the memory, the at least one processor configured to perform a method comprising;

collecting the first data;

correlating the first characteristics with the adoption results to generate a correlation result;

prior to installation of the first software on the second computer, collecting the second data; and

making a determination whether to install the first software on the second computer based on at least the second data and the correlation result.

19. The system of claim 18, wherein the collecting the second data is performed via an installer during installation of a second software on the second computer, and wherein the method further comprises in response to the determination, offering the first software as a bundle with the second software.

20. The system of claim 18, wherein the adoption results include a result selected from the group consisting of a paid upgrade by a user from a first version to a second version of the first software, use of the first software by a user for more than a predetermined time period, and a combination of the foregoing.