Predicting user interests

• US 9,633,312 B1
• Filed: 04/28/2015
• Issued: 04/25/2017
• Est. Priority Date: 02/02/2011
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Baluja, Shumeet, Covell, Michele
• Primary Examiner(s): Chaki, Kakali
• Assistant Examiner(s): Hanchak, Walter
• Application Number:
  US14/698,211
• Time in Patent Office: 728 Days
• Field of Search: None
• CPC Class Codes: G06F 16/2457 with adaptation to user needs
• CPC Class Codes: G06F 16/9038 Presentation of query results
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06N 3/00 Computing arrangements based on biological models
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for predicting user interests. In one aspect, a method includes training a prediction model to generate a category prediction of a next query from a set of queries, the category prediction specifying categories to which a next query belongs, the prediction model comprising a plurality of prediction functions that each generate a category prediction of a next query, wherein the prediction functions include two or more of a time-based prediction functions that generate a category prediction based on the category data and a difference of submission times of the queries, a rank-based prediction functions that generate a category prediction based on the category data and a rank order by which the queries were received, and a category-based prediction function that generates a category prediction based on the category data.

Claims

1. A method performed by data processing apparatus, the method comprising:

receiving queries from a user device, each received query being assigned to one or more topic-based categories and at least one of the received queries being assigned to multiple different topic-based categories;

identifying a set of topic-based categories for the user device, the set of topic-based categories comprising each topic-based category to which at least one of the received queries is assigned;

generating, using a plurality of prediction models, a category prediction of a next query to be received from the user device and that has not yet been received from the user device, the category prediction specifying at least one topic-based category to which the next query is predicted to belong, wherein;

the plurality of prediction models comprising at least two prediction models that each generate a respective category prediction based on the one or more topic-based categories to which each received query is assigned;

each of the plurality of prediction models using different criteria for generating the respective category predictions; and

at least one of the prediction models generates a respective category prediction based on a time at which at least one received query was received for each topic-based category in the set of topic-based categories and at least one of the prediction models generates a respective category prediction independent of the time at which at the least one received query was received for each topic-based category in the set of topic-based categories; and

providing, to the user device, data that cause a content item related to the at least one topic-based category to be presented at the user device.

2. The method of claim 1, wherein the prediction models comprise a time-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and a difference in submission times of the received queries.

3. The method of claim 1, wherein the prediction models comprise a category-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and irrespective of submission times of the received queries.

4. The method of claim 1, wherein the prediction models comprise a rank-based prediction model that generates a respective category prediction of the next query based on a rank order in which the received queries are ordered.

5. The method of claim 1, further comprising selecting the plurality of prediction models from a set of prediction models based on a domain from which a current query was received.

6. The method of claim 1, further comprising selecting the plurality of prediction models from a set of prediction models based on a type of the user device.

7. The method of claim 1, wherein each prediction model generates a score for each respective category prediction, the method further comprising:

weighting each score using a weight for the prediction model that generated the score, each prediction model having a different weight from each other prediction model; and

selecting the at least one topic-based category for use in identifying the content item based on the weighted scores for the category predictions.

20. The method of claim 5, wherein selecting the plurality of prediction models from a set of prediction models based on a domain from which a current query was received comprises:

selecting a first set of prediction models in response to the current query being received from a first domain category; and

selecting a second set of prediction models in response to the current query being received from a second domain category different from the first domain category, wherein the first set of prediction models is different from the second set of prediction models.

8. A system comprising:

a data processing apparatus; and

a memory storage apparatus in data communication with the data processing apparatus, the memory storage apparatus storing instructions executable by the data processing apparatus and that upon such execution cause the data processing apparatus to perform operations comprising;

receiving queries from a user device, each received query being assigned to one or more topic-based categories and at least one of the received queries being assigned to multiple different topic-based categories;

identifying a set of topic-based categories for the user device, the set of topic-based categories comprising each topic-based category to which at least one of the received queries is assigned;

generating, using a plurality of prediction models, a category prediction of a next query to be received from the user device and that has not yet been received from the user device, the category prediction specifying at least one topic-based category to which the next query is predicted to belong, wherein;

the plurality of prediction models comprising at least two prediction models that each generate a respective category prediction based on the one or more topic-based categories to which each received query is assigned;

each of the plurality of prediction models using different criteria for generating the respective category predictions; and

at least one of the prediction models generates a respective category prediction based on a time at which at least one received query was received for each topic-based category in the set of topic-based categories and at least one of the prediction models generates a respective category prediction independent of the time at which at the least one received query was received for each topic-based category in the set of topic-based categories; and

providing, to the user device, data that cause a content item related to the at least one topic-based category to be presented at the user device.

9. The system of claim 8, wherein the prediction models comprise a time-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and a difference in submission times of the received queries.

10. The system of claim 8, wherein the prediction models comprise a category-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and irrespective of submission times of the received queries.

11. The system of claim 8, wherein the prediction models comprise a rank-based prediction model that generates a respective category prediction of the next query based on a rank order in which the received queries are ordered.

12. The system of claim 8, wherein the operations further comprise selecting the plurality of prediction models from a set of prediction models based on a domain from which a current query was received.

13. The system of claim 8, wherein the operations further comprise selecting the plurality of prediction models from a set of prediction models based on a type of the user device.

14. The system of claim 8, wherein each prediction model generates a score for each respective category prediction, the operations further comprising:

weighting each score using a weight for the prediction model that generated the score, each prediction model having a different weight from each other prediction model; and

selecting the at least one topic-based category for use in identifying the content item based on the weighted scores for the category predictions.

15. A non-transitory computer storage medium encoded with a computer program, the program comprising instructions that when executed by data processing apparatus cause the data processing apparatus to perform operations comprising:

receiving queries from a user device, each received query being assigned to one or more topic-based categories and at least one of the received queries being assigned to multiple different topic-based categories;

identifying a set of topic-based categories for the user device, the set of topic-based categories comprising each topic-based category to which at least one of the received queries is assigned;

generating, using a plurality of prediction models, a category prediction of a next query to be received from the user device and that has not yet been received from the user device, the category prediction specifying at least one topic-based category to which the next query is predicted to belong, wherein;

the plurality of prediction models comprising at least two prediction models that each generate a respective category prediction based on the one or more topic-based categories to which each received query is assigned;

each of the plurality of prediction models using different criteria for generating the respective category predictions; and

at least one of the prediction models generates a respective category prediction based on a time at which at least one received query was received for each topic-based category in the set of topic-based categories and at least one of the prediction models generates a respective category prediction independent of the time at which at the least one received query was received for each topic-based category in the set of topic-based categories; and

providing, to the user device, data that cause a content item related to the at least one topic-based category to be presented at the user device.

16. The non-transitory computer storage medium of claim 15, wherein the prediction models comprise a time-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and a difference in submission times of the received queries.

17. The non-transitory computer storage medium of claim 15, wherein the prediction models comprise a category-based prediction model that generates a respective category prediction of the next query based on the one or more topic-based categories to which each received query is assigned and irrespective of submission times of the received queries.

18. The non-transitory computer storage medium of claim 15, wherein the prediction models comprise a rank-based prediction model that generates a respective category prediction of the next query based on a rank order in which the received queries are ordered.

19. The non-transitory computer storage medium of claim 15, wherein the operations further comprise selecting the plurality of prediction models from a set of prediction models based on a type of the user device.