티스토리 뷰
1. Rank
The rank of matrix in the number of pivots ($r(A) = \# $ of pivots)
- Also defiend as the maximum number of linearly independent columns (will be discussed later)
The rank of a matrix gives in a sense the true size of the matrix. Consider an $(m \times n)$ matrix $A$ whose rank is few than $m$. While $A \mathbf{x} = \mathbf{0}$ seems to have $m$ linear equations, $m-r$ out of $m$ equations are redundant in the sense that they can be derived from the rest of the equations.
From now on, consider $(m \times n)$ size matrix $A$.
Note : $r(A) \leq \min \{m, n\}$ (why? each column and row can have at most one pivot)
2. Full column Rank
full column rank : matrix $A$ whose $r(A) = n$
- $n$ pivot columns
- no free variables
- $N(A) = \{ \mathbf{0} \}$ ($N(A)$는 free variable의 선형 결합으로 표현되는데 free variable이 없으므로)
- $A \mathbf{x} = \mathbf{b}$ has $0$ or $1$ number of solution ($\mathbf{b}$가 column vector들의 span에 있는지 여부에 따름)
3. Full row Rank
full row rank : matrix $A$ whose $r(A) = m$
- $A \mathbf{x} = \mathbf{b}$ always solvable (see below)
- $C(A) = \mathbb{R}^{m}$ ($\mathbf{b} \in \mathbb{R}^{m}$인 어떤 $\mathbf{b}$에 대해서라도 $A \mathbf{x} = \mathbf{b}$가 solvable 하므로, $\mathbf{b} \in C(A)$라는 뜻이다. 그러려면 $C(A) = \mathbb{R}^{m}$ 이다.
- $n-m$ special solution(s)
4. $r(A)$ and Number of Solutions to $A \mathbf{x} = \mathbf{b}$
$r(A)$ | Number of solutions |
$r(A) = m \ \& \ r(A) = n$ | $1$ sol ($A$ : square and invertible) |
$r(A) = m \ \& \ r(A) < n$ (full row rank) | $\infty$ solutions ($n-r$ special solutions) |
$r(A) < m \ \& \ r(A) = n$ (full column rank) | $0$ or $1$ solution (no special solutions, $N(A) = \{\mathbf{0} \}$) |
$r(A) < m \ \& \ r(A) < n$ | $0$ or $\infty$ solutions ($x_{p}$(particular solution)의 존재 여부에 따름) |
5. Linear Independence
Vectors $\mathbf{v_{1}}, \mathbf{v_{1}}, ..., \mathbf{v_{n}}$ are said to be linearly independent if their linear combination $x_{1} \mathbf{v_{1}} + x_{2} \mathbf{v_{2}} + ... + x_{n} \mathbf{v_{n}} = \mathbf{0}$ only when $x_{i} = 0, \ \forall i$. Vectors that are not linearly independent are linearly dependent
Interpretation
- dependent 하다면 어떤 한 벡터는 나머지 다른 벡터들의 선형 결합으로 표현된다.
Rank and Independence
- Indeed, $r(A)$ is also defined as the maximum number of linearly independent columns in $A$
- It can be shown that row rank is equal to column rank for any matrix
- So, $r(A) = r(A^{T}) = \# $ of maximum linearly independent columns (rows)
6. Checking Linear Independence
Fact : Any set of $n$ vectors in $\mathbb{R}^{m}$ must be linearly dependent if $n > m$
(그 벡터들을 column으로 하는 행렬 $A$를 만들었을 때, $r(A)$의 최댓값이 $m$이기 때문이다. 이 말은 linearly independent한 column vector의 개수가 $m$이라는 뜻이므로 $n-m$개의 나머지 벡터들은 linearly dependent 할 수 밖에 없다)
How to check whether set of vectors are linearly independent or not?
Given : $\mathbf{v_{1}}, \mathbf{v_{1}}, ..., \mathbf{v_{n}} \in \mathbb{R}^{m}$
- If $n > m$, then dependent
- Else
- Construct a matrix $A = [\mathbf{v_{1}}, \mathbf{v_{1}}, ..., \mathbf{v_{n}} ]$
- Apply GE to $A$, and compute $r(A)$
- If $r(A) = n$, then independent. Ohterwise, dependent
'선형대수학' 카테고리의 다른 글
7. Basis and Dimension of Subspace (0) | 2024.02.29 |
---|---|
5. Subspaces (0) | 2024.02.25 |
4. Inverse Matrices (0) | 2024.02.20 |
3. Gaussian Elimination and LU Factorization (0) | 2024.02.18 |
2. Matrix Multiplication (0) | 2024.02.17 |