Skip to main content

Collaborative filtering concept

Short answer

Collaborative filtering recommends items by collecting preferences from similar users without needing item content. It builds a user-item matrix, computes neighbor similarity, and predicts missing ratings. For example, if Ram and Sita both liked films A and B, and Sita liked C, then C is recommended to Ram.

What do you mean by a collaborative filtering? Discuss the concept of collaborative filtering with suitable practical example.

Meaning

Collaborative filtering (CF) is a recommendation method that suggests items to a user based on the opinions or behaviour of other people with similar interests. The idea is "people who agreed in the past will agree in the future". It does not need to know the content of the items.

Concept

  • Data is stored as a user-item matrix of ratings or actions (most entries are empty).
  • The system finds users (or items) that are similar using a measure such as cosine similarity or Pearson correlation.
  • It predicts the missing rating as a weighted average of ratings by similar users, then recommends the highest predicted items.

Types

  • User-based CF: find users similar to the target user.
  • Item-based CF: find items similar to those the user liked (used by Amazon).
  • Model-based CF: learn a model, such as matrix factorization.

Practical example

Film AFilm BFilm C
Ram54?
Sita545
Hari121

Ram's ratings are close to Sita's and far from Hari's, so Sita is his neighbour. Sita rated C = 5, so the system predicts a high rating for Ram and recommends film C.

Advantages and problems

  • Advantages: works for any item type, finds unexpected items, improves with more data.
  • Problems: cold start, sparsity, scalability, popularity bias and privacy.

✓ Answer checked by Claude Opus 5.5 (high effort)

Related questions

All Information System chapter 7 questions →