Recommender System
An example of a network based recommender is Instagram’s suggested Instagrammer feature.
Description
Instagram can generate recommended users based on a number of factors such as followers, activity, location, engagement with others (i.e. comments), composition, and filtering options.
Reverse Engineering
Based on examination of how this is done and feedback from users who have made the suggested Instgrammer list the factors and algorithms used for determination of a suggested user are constantly evolving - but a few core characteristics remain as criteria for reverse engineering:
- followers, and number of edges or weighted distance as represented as nodes
- a ranked, or score for engagement with other users and a bias for temporal “freshness”
- content creation/publication with a temporal bias for “freshness”