Hands-On Recommendation Systems with Python

Singular-value decomposition

In Chapter 5, Getting Started with Data Mining Techniques, we mentioned that the math behind singular-value decomposition is well outside the scope of this book. However, let's try to gain an understanding of how it works from a layman's perspective.

Recall from Chapter 5, Getting Started with Data Mining Techniques, that PCA ( Principal Component Analysis) transforms an m × n matrix into n, m-dimensional vectors (called principal components) in such a way that each component is orthogonal to the next component. It also constructs these components in such a way that the first component holds the most variance (or information), followed by the second component, and so on.

Let's denote our ratings matrix as A. ...