Attribution of activity in multi-user settings

Patent metadata

Field	Value
Publication number	US-9818065-B2
Application number	US-201414206115-A
Country	US
Kind code	B2
Filing date	Mar 12, 2014
Priority date	Mar 12, 2014
Publication date	Nov 14, 2017
Grant date	Nov 14, 2017

Abstract

The claimed subject matter includes a system and method for attribution of search activity in multi-user settings. The method includes training a classifier to distinguish between machines that are single-user and multi-user based on activity logs of an identified machine. The identified machine is determined to be multi-user based on the classifier. A number of users is estimated for the identified machine. Activity of the users is clustered based on the number of users estimated. A similarity function is learned for the number of users estimated. The method also includes assigning new activity to one of the users based on the clustering, and the similarity function.

First claim

What is claimed is:

1. A method for attribution of activity in multi-user settings, the method comprising:
determining an identified machine is used by multiple users by using a classifier that is trained using activity logs of the identified machine;
estimating a number of the users for the identified machine by using an estimating regressor that is trained using the activity logs; and
assigning a new activity to one of the users by determining a similarity between the new activity and activities of a plurality of clusters, wherein the similarity is determined using a pair-wise similarity function that is learned by training a similarity regressor, and wherein each of the clusters represents activities of one of the users.

2. The method of claim 1 , the activity logs comprising logs of search activity.

3. The method of claim 2 , comprising making a recommendation for the one user based on the activity logs, the recommendation comprising one of a movie, book, song, video, or game.

4. The method of claim 1 , the activity logs comprising online activities.

5. The method of claim 1 , comprising presenting search results for the new activity based on the one user.

6. The method of claim 1 , comprising selecting features to classify search activity into clusters.

7. The method of claim 6 , comprising classifying search activity based on temporal features of the search activity.

8. The method of claim 1 , comprising allowing queries from multiple searchers on a shared machine to be hidden from other users.

9. The method of claim 1 , comprising generating metrics for each of the users of the identified machine.

10. The method of claim 1 , wherein assigning new activity comprises at least one of:
applying blind source separation methods to the activity;
Web site activity clustering;
fraud detection; or
diarization methods.

11. A system for attribution of search activity in multi-user settings, comprising:
a processing unit; and
a system memory, wherein the system memory comprises code configured to direct the processing unit to:
determine an identified machine is used by multiple users by using a classifier that is trained using activity logs of the identified machine;
estimate a number of the users for the identified machine by using an estimating regressor that is trained using the activity logs; and
assign a new activity to one of the users by determining a similarity between the new activity and activities of a plurality of clusters, wherein the similarity is determined using a pair-wise similarity function that is learned by training a similarity regressor, and wherein each of the clusters represents activities of one of the users.

12. The system of claim 11 , the system memory comprising code configured to direct the processing unit to make a recommendation for the one user.

13. The system of claim 12 , the recommendation comprising one of a movie, book, song, video, or game.

14. The system of claim 11 , the system memory comprising code configured to direct the processing unit to present a personalized advertisement for the one user.

15. The system of claim 11 , the system memory comprising code configured to direct the processing unit to present search results for the new activity based on the one user.

16. The system of claim 11 , the system memory comprising code configured to direct the processing unit to select features to classify search activity into clusters.

17. The system of claim 16 , the system memory comprising code configured to direct the processing unit to classify search activity based on temporal features of the search activity.

18. One or more computer-readable storage memory devices for storing computer-readable instructions, the computer-readable instructions attributing search activity in multi-user settings when executed by one or more processing devices, the computer-readable instructions comprising code configured to:
determine an identified machine is used by multiple users by using a classifier that is trained using activity logs of the identified machine;
estimate a number of the users for the identified machine by using an estimating regressor that is trained using the activity logs;
assign a new activity to one of the users by determining a similarity between the new activity and activities of a plurality of clusters, wherein the similarity is determined using a pair-wise similarity function that is learned by training a similarity regressor, and wherein each of the clusters represents activities of one of the users;
and
make a recommendation for the one user, the recommendation comprising one of a movie, book, song, video, or game.

19. The computer-readable storage memory devices of claim 18 , comprising code configured to direct the processing unit to present search results for the new activity based on the one user.

20. The computer-readable storage memory devices of claim 19 , comprising code configured to direct the processing devices to select features to classify search activity into clusters.

21. The method of claim 1 , wherein determining the identified machine is used by multiple users comprises estimating a number of searchers associated with an identifier of the identified machine.

Assignees

• Microsoft Corp
• Microsoft Technology Licensing Llc

Inventors

• White Ryen
• Awadallah Ahmed Hassan
• Singla Adish
• Horvitz Eric

Classifications

• G06N99/005 Primary

  Physics · mapped topic

• G06N20/00 Primary

  Machine learning · CPC title