Intermediate · Data
Item matrix
Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.
TL;DR. In recommender systems, a matrix containing latent feature representations (embeddings) for each item.
Technical Definition
In recommender systems, a matrix containing latent feature representations (embeddings) for each item.
How it works
The item matrix is a fundamental component in matrix factorization-based recommendation systems. It stores embedding vectors that capture hidden characteristics or latent factors for each item in the catalog. Each row typically represents an item, and each column corresponds to a latent feature, collectively encoding item properties that influence user preferences.
Related Concepts
- Embedding — A dense vector representation that captures semantic meaning, mapping discrete items like words into continuous mathematical space.