Recommender Systems 101: Collaborative Filtering vs Content Based?

80% of Netflix viewing, Amazon purchases, Spotify plays come from recommendations. Rec systems learn user preferences to suggest relevant items here’s how the two main approaches work,​Recommender Systems 101.

Connect With Us: WhatsApp

Collaborative filtering: “Users like you”

Core idea: Users who agreed in past → agree in future.

Matrix factorization:

• Users × Items → Ratings/Predictions
• Decompose to User factors × Item factors

Examples:

• Netflix: “Fans of A also watched B.”
• User‑based/User‑based: KNN on user similarity.
  Cold start: New users/items lack data.​

Content‑based: “You liked similar items”

Approach:

• Item features (genres, tags, descriptions).
• User profile = weighted history.
• Recommend items with high profile similarity (cosine/TFIDF).

Strengths: Handles new items, explainable (“because you liked X, Y, Z”).
**Weak: ** Limited to user’s past tastes (no serendipity).​

Connect With Us: WhatsApp

Hybrid: Best of both

• Winning pattern: Weighted/Stacking/ML hybrids.
• Netflix: 75% collaborative + content + context (time, device).

Evaluation:

• Offline: RMSE, NDCG, Recalk.
• Online: Click‑through, retention lift, revenue per user.​