Скачать или смотреть Linear Algebra — 2.2: Elimination Matrices

Every row operation in Gaussian elimination is secretly a matrix multiplication. This video builds elimination matrices from scratch — starting with two ways to view matrix-vector products, then constructing the specific matrices E₂₁ and E₃₂ that encode each elimination step, culminating in the equation E₃₂E₂₁A = U.

Key concepts covered:
• Column view: a matrix times a column vector produces a linear combination of the matrix's columns
• Row view: a row vector times a matrix produces a linear combination of the matrix's rows
• Why left-multiplication is the mechanism behind row operations
• Constructing an elimination matrix by modifying the identity — placing the negative of the multiplier in the appropriate position
• E₂₁ encodes "subtract 3 times row 1 from row 2" with entry -3 in position (2,1)
• E₃₂ encodes "subtract 2 times row 2 from row 3" with entry -2 in position (3,2)
• Verifying an elimination matrix works using the dot-product rule for matrix entries
• Chaining elimination matrices: E₃₂(E₂₁A) = U, where U is upper triangular
• Associativity: parentheses can be regrouped, so E₃₂E₂₁ collapses into a single elimination matrix
• Non-commutativity: E₃₂E₂₁ ≠ E₂₁E₃₂ — order matters in matrix multiplication
• Preview of the next topic: inverse matrices that undo elimination, taking U back to A

━━━━━━━━━━━━━━━━━━━━━━━━
SOURCE MATERIALS
The source materials for this video are from    • 2. Elimination with Matrices.