Collaborative Filtering Recommender

01 · The problem

Users needed relevant film suggestions from sparse rating data.

02 · Approach

1. 1

   Built a user-item matrix from 5,000 ratings across 200 users.

2. 2

   Used cosine similarity for item-to-item correlation.

3. 3

   Generated top-10 personalized suggestions with predicted rating scores.

03 · Outcome

Personalized top-10 recommendations with predicted scores per user.