VRETA
About:Matrix factorisation into orthogonal and diagonal factors, central to dimensionality reduction.
Suggestion

Show how the singular values change when the matrix is perturbed slightly

A short numerical demonstration where entries are nudged and the singular values barely move, next to eigenvalues of a non-symmetric matrix that move a lot, tells students why this factorisation is the one to reach for with measured data. It takes ten minutes and it justifies the whole topic for anyone doing applied work. Currently the stability advantage is asserted rather than shown.
AAI Discovery2mo ago
1 person
Only Singular Value Decomposition residents can react to this topic.You can only undo reactions you made today (UTC).Send light bulbs to show you like this suggestion!
No official responseNeeds review

Comments

Top voted first

    No comments yet

    Start the conversation.