Matrix Factors in AI & Artificial Intelligence

Traditional recommenders use visible features like 'Genre' or 'Director'. Matrix Factorization discovers Latent Factors—hidden characteristics that the model identifies through the patterns of user behavior. One factor might represent 'Movies with dry humor and slow pacing,' and another 'Explosive blockbusters with simple plots.' The model doesn't give these factors names, but it assigns a numerical value to every user and item for each factor, creating a dense map of interest.

Singular Value Decomposition (SVD) is the mathematical engine that performs this split. It takes the sparse Rating Matrix and factorizes it into three parts: a user matrix, a diagonal matrix of 'Singular Values' (the importance of each factor), and an item matrix. By keeping only the top factors (dimensionality reduction), we remove the 'Noise' (random ratings) and focus on the 'Signals' (real user trends). This is what made SVD the dominant algorithm in the Netflix Prize competition.

Once the matrices are factorized, predicting a rating is simple. To find out if User X will like Movie Y, we take the Dot Product of User X's factor vector and Movie Y's factor vector. If both have high weights for the same hidden factors, the resulting score will be high. This 'Compression' approach is extremely powerful for Sparse Data, as it allows the model to 'Infer' preferences even if a user has only rated a handful of items.

How do we find 'Hidden' features that users and items share? Matrix Factorization allows us to decompose a massive rating matrix into smaller, dense matrices of 'Latent Factors'.

We split the Rating Matrix (R) into two smaller matrices: Users (U) and Items (V). The product of these two matrices reconstructs the original ratings.

By discovering these factors, we can predict ratings for the 'empty' cells in our matrix. This was the key to winning the $1 Million Netflix Prize.

Checkpoint: What is a 'Latent Factor' in Matrix Factorization?

• →A visible tag (like 'Action')
• →A hidden characteristic discovered by the math that helps explain why certain users like certain items

SVD (Singular Value Decomposition) is the most famous factorization method. It reduces noise and identifies the strongest patterns in the data.

By mastering matrix factorization, you build engines that can see deeper than surface-level tags, capturing the subtle essence of human preference.

Checkpoint: Why is Matrix Factorization better for 'Sparse' data than simple KNN?

• →It uses less RAM
• →It compresses the sparse matrix into dense 'Factor' matrices, allowing for better predictions when data is missing

Factorization mastered! You've seen the hidden patterns. Ready to take it to the next level with Neural Collaborative Filtering?

Reconstruct a Real Rating from Latent Factors. Finish reconstructing a predicted rating as the dot product of a user's and an item's latent factor vectors.