Systems and methods for filtering page recommendations

1. A computer-implemented method comprising:

determining, by a computing system, a profile model for a page that is accessible through the computing system, the profile model describing one or more modal characteristics of users of the computing system that have fanned the page, the determining further comprising:

determining, by the computing system, at least one probability distribution for a profile setting, the probability distribution being constructed using values associated with the profile setting by the users that have fanned the page;

determining, by the computing system, the one or more modal characteristics based at least in part on the probability distribution;

providing, by the computing system, page recommendations that reference the page to a plurality of users of the computing system that specified at least one profile setting value that corresponds to a first modal characteristic in the one or more modal characteristics;

determining, by the computing system, that a threshold amount of the page recommendations resulted in at least one conversion by the users;

associating, by the computing system, the first modal characteristic with the page;

determining, by the computing system, that no profile setting values specified by a second user satisfy the first modal characteristic associated with the page; and

filtering, by the computing system, the page from being recommended to the second user.

2. The computer-implemented method of claim 1 , the method further comprising:

clustering, by the computing system, the values into one or more bins based at least in part on a semantic similarity or a string similarity.

3. The computer-implemented method of claim 1 , wherein the profile setting corresponds to at least one of the following characteristics: age, gender, gender preference, relationship status, occupation, workplace, education level, an institution of which the user is an alumni, religious affiliation, political affiliation, marital status, parental status, or causes supported by the user.

4. The computer-implemented method of claim 1 , wherein one or more of the values provided for the profile setting by the users that have fanned the page are weighted, wherein the weighting of a value specified by a user is based at least in part on a respective affinity between the user and the page.

5. The computer-implemented method of claim 1 , wherein determining the profile model for the page further comprises:

generating, by the computing system, a trained machine learning model for the page, the machine learning model being trained to predict whether profile setting values specified by a user correspond to the one or more modal characteristics of users that have fanned the page.

6. The computer-implemented method of claim 1 , the method further comprising:

determining, by the computing system, that the page should be recommended to a first user of the computing system, the determining further comprising:

determining, by the computing system, that the first user of the computing system has specified at least one profile setting value that corresponds to the first modal characteristic associated with the page; and

providing, by the computing system, at least one page recommendation to the first user that references the page.

7. A system comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to perform:

determining a profile model for a page that is accessible through the system, the profile model describing one or more modal characteristics of users of the system that have fanned the page, the determining further comprising:

determining at least one probability distribution for a profile setting, the probability distribution being constructed using values associated with the profile setting by the users that have fanned the page;

determining the one or more modal characteristics based at least in part on the probability distribution;

providing page recommendations that reference the page to a plurality of users of the system that specified at least one profile setting value that corresponds to a first modal characteristic in the one or more modal characteristics;

determining that a threshold amount of the page recommendations resulted in at least one conversion by the users;

associating the first modal characteristic with the page;

determining that no profile setting values specified by a second user satisfy the first modal characteristic associated with the page; and

filtering the page from being recommended to the second user.

8. The system of claim 7 , wherein the system further performs:

clustering the values into one or more bins based at least in part on a semantic similarity or a string similarity.

9. The system of claim 7 , wherein the profile setting corresponds to one of the following characteristics: age, gender, gender preference, relationship status, occupation, workplace, education level, an institution of which the user is an alumni, religious affiliation, political affiliation, marital status, parental status, or causes supported by the user.

10. The system of claim 7 , wherein the system further performs:

determining that the page should be recommended to a first user of the system, the determining further comprising:

determining that the first user of the system has specified at least one profile setting value that corresponds to the first modal characteristic associated with the page; and

providing at least one page recommendation to the first user that references the page.

11. The system of claim 7 , wherein the profile setting corresponds to at least one of the following characteristics: age, gender, gender preference, relationship status, occupation, workplace, education level, an institution of which the user is an alumni, religious affiliation, political affiliation, marital status, parental status, or causes supported by the user.

12. The system of claim 7 , wherein one or more of the values provided for the profile setting by the users that have fanned the page are weighted, wherein the weighting of a value specified by a user is based at least in part on a respective affinity between the user and the page.

13. The system of claim 7 , wherein the system further performs:

generating a trained machine learning model for the page, the machine learning model being trained to predict whether profile setting values specified by a user correspond to the one or more modal characteristics of users that have fanned the page.

14. A non-transitory computer-readable storage medium including instructions that, when executed by at least one processor of a computing system, cause the computing system to perform a method comprising:

determining a profile model for a page that is accessible through the computing system, the profile model describing one or more modal characteristics of users of the computing system that have fanned the page, the determining further comprising:

determining at least one probability distribution for a profile setting, the probability distribution being constructed using values associated with the profile setting by the users that have fanned the page;

determining the one or more modal characteristics based at least in part on the probability distribution;

providing page recommendations that reference the page to a plurality of users of the computing system that specified at least one profile setting value that corresponds to a first modal characteristic in the one or more modal characteristics;

determining that a threshold amount of the page recommendations resulted in at least one conversion by the users;

associating the first modal characteristic with the page;

determining that no profile setting values specified by a second user satisfy the first modal characteristic associated with the page; and

filtering the page from being recommended to the second user.

15. The non-transitory computer-readable storage medium of claim 14 , wherein the computing system further performs:

clustering the values into one or more bins based at least in part on a semantic similarity or a string similarity.

16. The non-transitory computer-readable storage medium of claim 14 , wherein the profile setting corresponds to one of the following characteristics: age, gender, gender preference, relationship status, occupation, workplace, education level, an institution of which the user is an alumni, religious affiliation, political affiliation, marital status, parental status, or causes supported by the user.

17. The non-transitory computer-readable storage medium of claim 14 , wherein the computing system further performs:

determining that the page should be recommended to a first user of the computing system, the determining further comprising:

determining that the first user of the computing system has specified at least one profile setting value that corresponds to the first modal characteristic associated with the page; and

providing at least one page recommendation to the first user that references the page.

18. The non-transitory computer-readable storage medium of claim 14 , wherein the profile setting corresponds to at least one of the following characteristics: age, gender, gender preference, relationship status, occupation, workplace, education level, an institution of which the user is an alumni, religious affiliation, political affiliation, marital status, parental status, or causes supported by the user.

19. The non-transitory computer-readable storage medium of claim 14 , wherein one or more of the values provided for the profile setting by the users that have fanned the page are weighted, wherein the weighting of a value specified by a user is based at least in part on a respective affinity between the user and the page.

20. The non-transitory computer-readable storage medium of claim 14 , wherein the computing system further performs:

generating a trained machine learning model for the page, the machine learning model being trained to predict whether profile setting values specified by a user correspond to the one or more modal characteristics of users that have fanned the page.