How does Facebook suggests friends? | Graph theory| Learn with me Jeya

Summary

This short explainer breaks down how Facebook’s “People You May Know” suggestions work. It highlights four core signals—profile data (work, school, location), user activity (groups and tags), uploaded contacts, and mutual friends—and shows how they combine to produce friend recommendations. The video visualizes the network as a graph (users as nodes, friendships as edges) and explains the mutual-friend approach using a breadth-first search to find two-hop nodes (friends-of-friends), then ranking candidates by their number of common neighbors. Clear visuals and a simple graph example make the algorithm easy to grasp, ending with a friendly like-and-subscribe call-to-action.

Takeaways

• 😀 Facebook is one of the most popular social networking sites, connecting millions of people globally.
• 😀 The 'People You May Know' feature on Facebook suggests potential friends based on various factors.
• 😀 The friend suggestion algorithm relies on four key factors: profile data, Facebook activities, contacts, and mutual friends.
• 😀 Profile data includes your workplace, education, location, and other personal information to make suggestions.
• 😀 Facebook tracks your activities such as group memberships and people tagged in posts and photos to suggest friends.
• 😀 Contacts are also a factor in the algorithm, including people you have uploaded or have uploaded your contact information.
• 😀 Mutual friends play a major role in friend suggestions, using the common neighbor algorithm.
• 😀 The common neighbor algorithm treats the entire social network as a graph, where users are nodes and friendships are edges.
• 😀 To find friend suggestions, Facebook searches for nodes (people) connected by mutual friends, using breadth-first search (BFS).
• 😀 Nodes are ranked based on the number of common neighbors they share with the user, prioritizing those with the most mutual connections.
• 😀 As a result, friend suggestions are made in order of common neighbors, showing the most likely connections first.

Q & A

What is the purpose of the 'People You May Know' section on Facebook?

The 'People You May Know' section suggests potential friends to users, helping them connect with new people based on various factors like shared data, activities, and mutual friends.

What are the four key factors that influence Facebook's friend suggestion algorithm?

The four key factors are: profile data (workplace, education, location), Facebook activities (groups joined, people tagged in photos/posts), contacts (people you've uploaded or who've uploaded your contact), and mutual friends (common friends between users).

How does profile data contribute to Facebook's friend suggestions?

Profile data such as workplace, education, and location is used by Facebook to suggest friends who have similar or matching details, increasing the likelihood of a connection.

What role do Facebook activities play in friend suggestions?

Facebook activities, such as joining groups and being tagged in posts or photos, help the algorithm identify potential connections based on shared interests or interactions.

How do contacts affect Facebook's friend suggestion algorithm?

Contacts influence the algorithm by including people you've uploaded to Facebook or those who have uploaded your contact, suggesting friends from these networks.

What is the 'mutual friends' factor in the friend suggestion algorithm?

The mutual friends factor suggests new friends based on the common friends both users share, leveraging the idea that people with shared connections are more likely to connect.

What is the Common Neighbor Algorithm and how does it work?

The Common Neighbor Algorithm is used to find potential friends based on shared connections. It works by identifying mutual friends between users and ranking them according to the number of common neighbors.

How is a social network visualized in the context of the Common Neighbor Algorithm?

A social network is visualized as a graph, where users are represented as nodes, and friendships are edges connecting those nodes. The algorithm looks for nodes that are connected through common neighbors.

What does 'breadth-first search' mean in the context of this algorithm?

Breadth-first search is a method used to explore the network graph. It helps identify nodes (users) that are one step away from a given node (user) by examining their immediate neighbors.

Why does Facebook rank potential friends based on the number of common neighbors?

Facebook ranks potential friends based on common neighbors because users with more shared connections are more likely to have mutual interests or ties, making the suggestion more relevant.

How does the algorithm determine which suggested friend to display first?

The algorithm ranks suggested friends by the number of common neighbors. The friend with more mutual connections is displayed higher, making it more likely for users to connect with people they have more in common with.