- independent .VS. dependent
给定向量u, v, w,如果x1*u+x2*v+x3*w = 0当且仅当x1=0,x2=0,x3=0时才成立,则称u,v,w是independent
Independent columns: Ax = 0 has one solution. A is an inveritlbe matrix
Depedent columns: Ax = 0 has many solutions. A is a singular matrix
- Gauss-Jordan Elimination
可以手动的求逆,[A I]通过初等变换得到[I B],B就是A的逆
- LU factorization
L是下三角矩阵,U是上三角矩阵,U的对角元素是pivot
(E32E31E21)A = U 变成 A = (E21-1 E31-1 E32-1 )U 即 A=LU
简单来说,L就是把原来初等矩阵对焦元素下方的符号变了
- LDU factorization
也就是对LU中的U进一步分解提出来个D={dii =di },把U每一行除以di
- LU factorization to solve equation
把one square system变成two triangular systems
Solve Ax = b -----> Solve Lc = b and then solve Ux = c
- Transpose
Ax combines the columns of A while xTAT combines the rows of AT
- Symmetric matrix
如果对称矩阵(A=AT),对它做LDU分解,则U=LT
- Permuation matrix
置换矩阵(使矩阵行交换),元素只有0和1,对于n阶矩阵,只有n!个置换矩阵。
1. 如果P是置换矩阵,则P-1也是置换矩阵,同时P-1=PT
2. PA = LU
- Fundamental Theorem of Linear Algebra, Part I
In Rn the row space and nullspace have dimensions r and n-r (total n).
In Rm the column space and left nullspace have dimensions r and m-r (total m).
- Fundamental Theorem of Linear Algebra, Part II
N(A) is the orthogonal complement of the row space C(AT) (in Rn)
N(AT) is the orthogonal complement of the column space C(A) (in Rm)
- Projections
The projection of b onto the subspace is p: p = A x = A(ATA)-1ATb
Th n by n projection matrix that produces p = Pb: P = A(ATA)-1AT
- Gram-Schmidt --- a way to create orthonormal vectors
Start with three independent vectors a, b, c
A = a, B = b - (ATb)/(ATA) A, C = c - (ATc)/(ATA) A - (BTc)/(BTB)B
q1 = A/||A||, q2 = B/||B||, q3 = C//||C||
- QR factorization
[a b c] = [q1 q2 q3][q1Ta q1Tb q1Tc; 0 q2Tb q2Tc; 0, 0, q3Tc]
- The Properties of the Determinant
The determinant of the n by n identity matrix is 1.
The determinant changes sign when two rows are exchanged.
The determinant is a linear function of each row separately.
If two rows of A are equal, then det A = 0.
Substracting a multiple of one row from another row leaves det A unchanged.
A matrix with a row of zeros has det A = 0
If A is triangular then det A = a11a22...ann = product of diagonal entries.
If A is singular then det A = 0. If A is invertible then det A doesn't equal 0 .
The determinant of AB is det A times det B: |AB| = |A||B|.
The transpose AT has the same determinant as A.
- Eigenvalues Ax = λx
Each column of A adds to 1, so λ = 1 is an eigenvalue.
A is singular, so λ = 0 is an eigenvalue.
A is symmetric, so its eigenvectors are perpendicular.
λ1+λ2+...+λn = trace = a11 + a22 + ... + ann.
λ1*λ2*...*λn = determinant(A).