Systems and methods to determine location recommendations

1. A computer-implemented method comprising:

receiving, by a computing system, ratings, provided by at least a user and a set of other users through a social networking system, for a plurality of locations associated with a location type, wherein receiving the ratings comprises:

receiving a rating for at least a first location from the user and ratings for at least the first location from the set of other users; and

receiving ratings for a second location from the set of other users;

developing, by the computing system, a personalized model for the user, based on attributes associated with the location type and the ratings for the plurality of locations wherein the developing comprises:

calculating a pairwise difference between the rating for at least the first location from the user and each rating for at least the first location from each user in the set of other users to generate difference values, wherein the difference values are weighted by an exponential decay function based on a similarity between the rating from the user and each rating from each user in the set of other users;

generating a confidence interval based at least in part on the weighted difference values;

calculating an expected rating for the second location based on the ratings for the second location from the set of other users; and

applying the confidence interval to the expected rating, wherein the confidence interval is associated with an accuracy of the expected rating;

providing, by the computing system, the second location as a recommendation for the user based on the personalized model based at least in part on whether the confidence interval satisfies a threshold accuracy;

determining, by the computing system, that an actual rating received from the user for the second location does not fall within the confidence interval; and

developing, by the computing system, the personalized model further for the user based on the actual rating and the ratings for the second location from the set of other users.

2. The computer-implemented method of claim 1 , wherein the location type relates to at least one of: restaurants, stores, schools, bars, companies, congregations, or destinations.

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

for each respective location of the plurality of locations, determining attribute values of each attribute associated with the location type.

4. The computer-implemented method of claim 3 , wherein at least a portion of the attribute values are determined based at least in part on user data maintained by the social networking system.

5. The computer-implemented method of claim 3 , wherein the developing the personalized model for the user further comprises:

correlating the ratings for the plurality of locations with the attribute values for each respective location of the plurality of locations;

identifying desired attribute values that are desired by the user based on the correlating; and

developing the personalized model to reflect the desired attribute values that are desired by the user.

6. The computer-implemented method of claim 5 , wherein the developing the personalized model for the user further comprises:

assigning a weight to each attribute value to reflect an importance of each attribute to the user; and

developing the personalized model to reflect the assigned weight.

7. The computer-implemented method of claim 1 , wherein generating the confidence interval comprises combining at least one of: the difference values or the weighted difference values.

8. The computer-implemented method of claim 1 , wherein the developing the personalized model further for the user further comprises:

developing the personalized model further for the user based on additional ratings for additional locations from the user and the set of other users.

9. 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:

receiving ratings, provided by at least a user and a set of other users through a social networking system, for a plurality of locations associated with a location type, wherein receiving the ratings comprises:

receiving a rating for at least a first location from the user and ratings for at least the first location from the set of other users; and

receiving ratings for a second location from the set of other users;

developing a personalized model for the user, based on attributes associated with the location type and the ratings for the plurality of locations, wherein the developing comprises:

calculating a pairwise difference between the rating for at least the first location from the user and each rating for at least the first location from each user in the set of other users to generate difference values, wherein the difference values are weighted by an exponential decay function based on a similarity between the rating from the user and each rating from each user in the set of other users;

generating a confidence interval based at least in part on the weighted difference values;

calculating an expected rating for the second location based on the ratings for the second location from the set of other users; and

applying the confidence interval to the expected rating, wherein the confidence interval is associated with an accuracy of the expected rating;

providing the second location as a recommendation for the user based on the personalized model based at least in part on whether the confidence interval satisfies a threshold accuracy;

determining that an actual rating received from the user for the second location does not fall within the confidence interval; and

developing the personalized model further for the user based on the actual rating and the ratings for the second location from the set of other users.

10. The system of claim 9 , wherein the location type relates to at least one of: restaurants, stores, schools, bars, companies, congregations, or destinations.

11. The system of claim 9 , wherein the instructions further cause the system to perform:

for each respective location of the plurality of locations, determining attribute values of each attribute associated with the location type.

12. The system of claim 11 , wherein the developing the personalized model for the user further comprises:

correlating the ratings for the plurality of locations with the attribute values for each respective location of the plurality of locations;

identifying desired attribute values that are desired by the user based on the correlating; and

developing the personalized model to reflect the desired attribute values that are desired by the user.

13. The system of claim 12 , wherein the developing the personalized model for the user further comprises:

assigning a weight to each attribute value to reflect an importance of each attribute to the user; and

developing the personalized model to reflect the assigned weight.

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:

receiving ratings, provided by at least a user and a set of other users through a social networking system, for a plurality of locations associated with a location type, wherein receiving the ratings comprises:

receiving a rating for at least a first location from the user and ratings for at least the first location from the set of other users; and

receiving ratings for a second location from the set of other users;

developing a personalized model for the user, based on attributes associated with the location type and ratings for the plurality of locations, wherein the developing comprises:

calculating a pairwise difference between the rating for at least the first location from the user and each rating for at least the first location from each user in the set of other users to generate difference values, wherein the difference values are weighted by an exponential decay function based on a similarity between the rating from the user and each rating from each user in the set of other users;

generating a confidence interval based at least in part on the weighted difference values;

calculating an expected rating for the second location based on the ratings for the second location from the set of other users; and

applying the confidence interval to the expected rating, wherein the confidence interval is associated with an accuracy of the expected rating;

providing the second location as a recommendation for the user based on the personalized model based at least in part on whether the confidence interval satisfies a threshold accuracy;

determining that an actual rating received from the user for the second location does not fall within the confidence interval; and

developing the personalized model further for the user based on the actual rating and the ratings for the second location from the set of other users.

15. The non-transitory computer-readable storage medium of claim 14 , wherein the location type relates to at least one of: restaurants, stores, schools, bars, companies, congregations, or destinations.

16. The non-transitory computer-readable storage medium of claim 14 , wherein the method further comprises:

for each respective location of the plurality of locations, determining attribute values of each attribute associated with the location type.

17. The non-transitory computer-readable storage medium of claim 16 , wherein the developing the personalized model for the user further comprises:

correlating the ratings for the plurality of locations with the attribute values for each respective location of the plurality of locations;

identifying desired attribute values that are desired by the user based on the correlating; and

developing the personalized model to reflect the desired attribute values that are desired by the user.

18. The non-transitory computer-readable storage medium of claim 17 , wherein the developing the personalized model for the user further comprises:

assigning a weight to each attribute value to reflect an importance of each attribute to the user; and

developing the personalized model to reflect the assigned weight.