The user recommendation system provides users with a means to discover like-minded users on the site with whom to connect. User recommendations are personalized for each user based on behavior and interactions on the site, including liking and rating content.
The user recommendation system will find similar users to the accessing user and recommend them as potential friends. The accessing user can then choose to request their friendship, follow the user, or ignore the recommendation. Users the accessing user is already friends with, is following, or has ignored will not be recommended to him/her.
The user recommendation system calculates recommendations in the Job Scheduler during off-peak hours. The system runs two different jobs for user recommendations. One, Telligent.Evolution.CoreServices.Recommendations.UserRecommendationCalculationJob, is run once a week at 2 A.M and updates recommendations for all users. The second, Telligent.Evolution.CoreServices.Recommendations.UserRecommendationCalculationForScheduledUsersJob, is run six days a week at 2 A.M and only updates recommendations for users that have visited the site in the last day. This is to avoid unnecessary recalculations while still keeping the recommendations up to date.
The jobs calculate 20 recommendations for each user at a time. If a user chooses to friend, follow, or ignore all 20 recommendations in a given day, that user will not receive any more user recommendations until the following day after the job has found more recommendations.
As mentioned earlier, users can ignore recommendations in the UI to not have that recommendation be made again. The X in the following image is how the user ignores the recommendation.
The user recommendation system is not currently open for extensibility or modification but recommendations can be retrieved via the APIs.