What is user-based collaborative filtering?
Updated May 16, 2026
Short answer
It recommends items based on similar users' preferences.
Deep explanation
User-based collaborative filtering finds users with similar behavior patterns and recommends items liked by those neighbors. Similarity is computed using metrics like cosine similarity or Pearson correlation over user-item interaction vectors.
Real-world example
Friend-based recommendations in social media platforms.
Common mistakes
- Not handling sparse user interaction matrices properly.
Follow-up questions
- Why is it less scalable?
- What is neighborhood size?