Systems and methods for generating scalable personalized recommendations based on dynamic and customized content selections and modeling of the content selections

1. A device, comprising:

one or more processors configured to:

maintain a plurality of recommendation configurations, wherein each respective recommendation configuration includes:

a respective set of selector criteria based on which the respective recommendation configuration is selected from the plurality of recommendation configurations,

a respective set of pre-boost parameters specifying criteria based on which content can be selected from a plurality of content and can further be prioritized,

a respective set of models to apply to content selected and prioritized according to the respective set of pre-boost parameters, wherein different recommendation configurations include different sets of models;

receive a request comprising a query and an identifier that identifies a requestor issuing the request;

compare the query and the identifier of the requestor to the respective sets of selector criteria associated with the plurality of recommendation configurations;

select a particular recommendation configuration based on comparing the query and the identifier of the requestor to the respective sets of selector criteria;

retrieve a set of content from a plurality of accessible content based on each content in the set of content satisfying the query and the set of pre-boost parameters included in the particular recommendation configuration;

prioritize the retrieved set of content in a first order based on the set of pre-boost parameters included in the particular recommendation configuration;

apply the respective set of models, associated with the particular recommendation configuration, to the set of content; and

present recommended content in a second order, that is different than the first order, based on applying the respective set of models, associated with the particular recommendation configuration that was selected based on the comparing, to the retrieved set of content.

2. The device of claim 1 ,

wherein the particular recommendation configuration specifies content-based filtering of the set of content against a first event, and collaborative filtering of the set of content against a second event, and

wherein applying the respective set of models comprises:

modeling a first set of relevance scores based on the content-based filtering of the set of content against the first event; and

modeling a second set of relevance scores based on the collaborative filtering of the set of content against the second event.

3. The device of claim 1 , wherein the one or more processors are further configured to:

evaluate the set of models according to an interleaving technique specified in the particular recommendation configuration;

rank the set of content based on results from evaluating the set of models; and

determine the second order for the recommended content based on the ranking of the set of content.

4. The device of claim 1 , wherein the one or more processors are further configured to:

obtain one or more preferences of the requestor;

modify at least one parameter of the one or more parameters using a particular preference from the one or more preferences; and

query a content repository for the set of content using the query modified with the particular preference.

5. The device of claim 1 , wherein the one or more processors are further configured to:

obtain one or more preferences of the requestor; and

prioritize selection of content from the plurality of accessible content that satisfies the query and that further satisfies the one or more preferences that are not specified in the query.

6. The device of claim 1 , wherein selecting the particular recommendation configuration comprises:

selecting a first recommendation configuration, including a first set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a first time; and

selecting a second recommendation configuration, including a second set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a second time that is different from the first time.

7. The device of claim 1 ,

wherein selector criteria for a first recommendation configuration of the plurality of recommendation configurations includes a first content type, and

wherein selector criteria for a second recommendation configuration of the plurality of recommendation configurations includes a second content type.

8. A non-transitory computer-readable medium, storing a plurality of processor executable instructions to:

maintain a plurality of recommendation configurations, wherein each respective recommendation configuration includes:

a respective set of selector criteria based on which the respective recommendation configuration is selected from the plurality of recommendation configurations,

a respective set of pre-boost parameters specifying criteria based on which content can be selected from a plurality of content and can further be prioritized,

a respective set of models to apply to content selected and prioritized according to the respective set of pre-boost parameters, wherein different recommendation configurations include different sets of models;

receive a request comprising a query and an identifier that identifies a requestor issuing the request;

compare the query and the identifier of the requestor to the respective sets of selector criteria associated with the plurality of recommendation configurations;

select a particular recommendation configuration based on comparing the query and the identifier of the requestor to a the respective sets of selector criteria;

retrieve a set of content from a plurality of accessible content based on each content in the set of content satisfying the query and the set of pre-boost parameters included in the particular recommendation configuration;

prioritize the retrieved set of content in a first order based on the set of pre-boost parameters included in the particular recommendation configuration;

apply the respective set of models, associated with the particular recommendation configuration, to the set of content; and

present recommended content in a second order, that is different than the first order, based on applying the respective set of models, associated with the particular recommendation configuration that was selected based on the comparing, to the retrieved set of content.

9. The non-transitory computer-readable medium of claim 8 ,

wherein the particular recommendation configuration specifies content-based filtering of the set of content against a first event, and collaborative filtering of the set of content against a second event; and

wherein applying the respective set of models comprises:

modeling a first set of relevance scores based on the content-based filtering of the set of content against the first event; and

modeling a second set of relevance scores based on the collaborative filtering of the set of content against the second event.

10. The non-transitory computer-readable medium of claim 8 , wherein the plurality of processor-executable instructions further include instructions to:

evaluate the set of models according to an interleaving technique specified in the particular recommendation configuration;

rank the set of content based on results from evaluating the set of models; and

determine the second order for the recommended content based on the ranking of the set of content.

11. The non-transitory computer-readable medium of claim 8 , wherein the plurality of processor-executable instructions further include instructions to:

obtain one or more preferences of the requestor;

modify at least one parameter of the one or more parameters using a particular preference from the one or more preferences; and

query a content repository for the set of content using the query modified with the particular preference.

12. The non-transitory computer-readable medium of claim 8 , wherein the plurality of processor-executable instructions further include instructions to:

obtain one or more preferences of the requestor; and

prioritize selection of content from the plurality of accessible content that satisfies the query and that further satisfies the one or more preferences that are not specified in the query.

13. The non-transitory computer-readable medium of claim 8 , wherein selecting the particular recommendation configuration comprises:

selecting a first recommendation configuration, including a first set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a first time; and

selecting a second recommendation configuration, including a second set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a second time that is different from the first time.

14. A method, comprising:

maintaining a plurality of recommendation configurations, wherein each respective recommendation configuration includes:

a respective set of selector criteria based on which the respective recommendation configuration is selected from the plurality of recommendation configurations,

a respective set of pre-boost parameters specifying criteria based on which content can be selected from a plurality of content and can further be prioritized,

a respective set of models to apply to content selected and prioritized according to the respective set of pre-boost parameters, wherein different recommendation configurations include different sets of models;

receiving a request comprising a query and an identifier that identifies a requestor issuing the request;

comparing the query and the identifier of the requestor to the respective sets of selector criteria associated with the plurality of recommendation configurations;

selecting a particular recommendation configuration based on comparing the query and the identifier of the requestor to the respective sets of selector criteria;

retrieving a set of content from a plurality of accessible content based on each content in the set of content satisfying the query and the set of pre-boost parameters included in the particular recommendation configuration;

prioritizing the retrieved set of content in a first order based on the set of pre-boost parameters included in the particular recommendation configuration;

applying the respective set of models, associated with the particular recommendation configuration, to the set of content; and

presenting recommended content in a second order, that is different than the first order, based on applying the respective set of models, associated with the particular recommendation configuration that was selected based on the comparing, to the retrieved set of content.

15. The method of claim 14 , further comprising:

obtaining one or more preferences of the requestor;

modifying at least one parameter of the one or more parameters using a particular preference from the one or more preferences; and

querying a content repository for the set of content using the query modified with the particular preference.

16. The method of claim 14 , further comprising:

obtaining one or more preferences of the requestor; and

prioritizing selection of content from the plurality of accessible content that satisfies the query and that further satisfies the one or more preferences that are not specified in the query.

17. The method of claim 14 , wherein selecting the particular recommendation configuration comprises:

selecting a first recommendation configuration, including a first set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a first time; and

selecting a second recommendation configuration, including a second set of models, as the particular recommendation configuration in response to one or more of the request being issued by the requestor at a second time that is different from the first time.

18. The method of claim 14 , wherein obtaining the particular recommendation configuration comprises:

wherein selector criteria for a first recommendation configuration of the plurality of recommendation configurations includes a first content type, and

wherein selector criteria for a second recommendation configuration of the plurality of recommendation configurations includes a second content type.

19. The method of claim 14 ,

wherein the particular recommendation configuration specifies content-based filtering of the set of content against a first event, and collaborative filtering of the set of content against a second event, and

wherein applying the respective set of models comprises:

modeling a first set of relevance scores based on the content-based filtering of the set of content against the first event; and

modeling a second set of relevance scores based on the collaborative filtering of the set of content against the second event.

20. The method of claim 14 , further comprising:

evaluating the set of models according to an interleaving technique specified in the particular recommendation configuration;

ranking the set of content based on results from evaluating the set of models; and

determining the second order for the recommended content based on the ranking of the set of content.