Why is matrix rank important in machine learning models?

Rank determines redundancy and expressive capacity of features.

Low rank implies redundant features and reduced expressive power. High rank indicates richer representation but may increase overfitting risk.