Matrix Rank — What It Means and How to Find It
Learn what matrix rank represents, how to calculate it using row echelon form, and why it matters for solving linear systems and understanding transformations.
Detailed Explanation
Matrix Rank
The rank of a matrix is the number of linearly independent rows (equivalently, linearly independent columns). It tells you the dimension of the image (column space) of the linear transformation represented by the matrix.
Finding the Rank
The most common method is to reduce the matrix to Row Echelon Form (REF) and count the number of non-zero rows:
A = | 1 2 3 | REF = | 1 2 3 |
| 2 4 6 | | 0 0 0 |
| 3 5 7 | | 0 0 2 |
After rearranging, there are 2 non-zero rows, so rank(A) = 2.
Properties
- rank(A) <= min(m, n) for an m x n matrix
- rank(A) = rank(A^T) (row rank equals column rank)
- rank(AB) <= min(rank(A), rank(B))
- rank(A) + nullity(A) = n (Rank-Nullity Theorem)
What Rank Tells You
For an m x n matrix A and the system Ax = b:
- Full column rank (rank = n): The system has at most one solution
- Full row rank (rank = m): The system is consistent for every b
- Full rank (rank = min(m,n)): Maximum rank possible
- Rank deficient (rank < min(m,n)): The matrix has dependent rows/columns
Rank and Determinant
A square n x n matrix has:
- rank = n if and only if det(A) != 0 (invertible)
- rank < n if and only if det(A) = 0 (singular)
Use Case
Matrix rank is fundamental in data science for dimensionality reduction (PCA relies on the rank of the covariance matrix). In control theory, the rank of the controllability matrix determines if a system is controllable. In computer vision, the rank of the fundamental matrix constrains epipolar geometry between two camera views.