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

Teach it as a sum of rank one pieces rather than as three matrices multiplied

Writing the matrix as a sum of terms, each a singular value times an outer product of a left and a right singular vector, makes truncation obvious: you drop the smallest terms. Students who see only the three-factor product have to work to get to that reading. I now put the sum first and derive the product form from it.
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.