Agent skill · research science · secondsky
recommendation-engine
Build recommendation systems with collaborative filtering, matrix factorization, hybrid approaches. Use for product recommendations, personalization, or encountering cold start, sparsity, quality evaluation issues.
Why this skill is useful
Includes executable scripts for collaborative filtering and matrix factorization that extend the AI's capabilities in building recommendation systems.
What it needs
About 5k tokens when loaded. Last updated 2026-08-06. 204 stars on the source repository.
What this skill does
Recommendation Engine Build recommendation systems for personalized content and product suggestions. Recommendation Approaches Approach How It Works Pros Cons ---------- -------------- ------ ------ Collaborative User-item interactions Discovers hidden patterns Cold start Content-based Item features Works for new items Limited discovery Hybrid Combines both Best of both Complex Collaborative Filtering Matrix Factorization (SVD) Hybrid Recommender Evaluation Metrics Precision@K, Recall@K NDCG (ranking quality) Coverage (catalog diversity) A/B test conversion rate Cold Start Solutions New users: Popular items, onboarding preferences, demographic-based New items: Content-based bootstrapping, active learning Exploration strategies: ε-greedy, Thompson sampling bandits Quick Start: Build a Recommender in 5 Steps Known Issues Prevention 1. Popularity Bias Problem: Recommending only popular items, ignoring long tail. Reduces diversity and serendipity. Solution: Balance popularity with personalization, apply re-ranking for diversity: 2. Data Sparsity (Matrix >99% Empty) Problem: Collaborative filtering fails when most users have rated <1% of items. Solution: Use matrix factorization (SVD, ALS) instead of memory-based CF: 3. Cold Start Without Fallback Problem: Recommender crashes or returns empty results for new users/items. Solution: Always implement fallback chain: 4. Not Excluding Already-Interacted Items Problem: Recommending items user already purchased/viewed wastes recommendation slots. Solution: Always filter interacted items: 5. Ignoring Implicit Feedback Confidence Problem: Treating all clicks/views equally. 1 view ≠ 100 views. Solution: Weight by interaction strength (view count, watch time, etc.): 6. Not Evaluating Ranking Quality (Using Only Accuracy) Problem: High prediction accuracy (RMSE) doesn't mean good top-K recommendations. Solution: Use ranking metrics (NDCG, MAP@K): 7. …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills secondsky/recommendation-engine