Home » Developer Resources » Resources: How to Implement Friend Suggestion

Resources: How to Implement Friend Suggestion

Looking to create a friend suggestion system for your social network? Or perhaps you’re just curious about Facebook’s uncanny ability at suggesting “friends” that you actually know? In this resource guide, you’ll find articles, videos, and tutorials on almost everything you need to know about friend recommendation system commonly used by social networks such as Facebook.

Social Network: Suggesting Friends

1. PHP Friend Suggestion System

The previous owner of this domain produced a solid article on how to create a friend suggestion system using PHP, MySQL, and jQuery.
It’s complete with code examples, live demo, and diagrams. The article is a little dated now but the concepts laid out by the author are still applicable and I believe the code is still up to today’s standard. There’s no machine learning involved and that means easier to maintain and faster development time.

Unfortunately, the article is no longer live, but for those of you who are interested, check out this article on Archive.org.

See it here

2. Non-Technical Friend Recommendation Overview

How does Facebook, the most popular social network in the world, take care of friend suggestions?

These articles provide an in-depth non-technical overview of the system that Facebook uses for recommending friends.

  • inosocial: How does Facebook suggest friends? Read it here.
  • HTCW: How Does Facebook Friend Suggestion Works. Read it here.

3. Machine Learning & Friend Recommendation

Machine learning is used in many recommendation systems. Most famously, movie recommendation and book recommendation. How Netflix does know what show you’re most likely to watch? The answer? Data mining and machine learning.

Personally, I think machine learning is an overkill for a friend recommendation system similar to Facebook or LinkedIn. In such systems, you can simply suggest friends or connections based on common connections, schools, employers, and get a pretty good accuracy. Machine learning might provide a marginal improvement over a non-AI solution, but developing and maintaining such a system is just not worth the effort.

That said, here are some pretty good research papers regarding machine learning in friend suggestion you should check out.

  • Link Prediction across Heterogeneous Social Networks: A Survey. See it here.
  • Facebook Friend Suggestion. See it here.
  • Friend Recommendation using Unsupervised Machine Learning. See it here.
  • datacamp: Beginner Tutorial: Recommender Systems in Python. See it here.

Lastly, if you’re new and you’re interested in machine learning, this machine learning course on Coursera is a good place to start.