Finding A Basis Of A Subspace

Finding a basis for a subspace is one of the most fundamental skills in linear algebra. It bridges the gap between abstract vector spaces and concrete computation, allowing us to describe infinite sets of vectors using a finite, manageable list. Whether you are working with the column space of a matrix, the solution set of a homogeneous system, or a subspace defined by linear constraints, the core objective remains the same: identify a set of vectors that is both linearly independent and spans the subspace.

Worth pausing on this one.

Understanding the Core Concepts

Before diving into algorithms, Make sure you solidify the definitions. It matters. A subspace $W$ of a vector space $V$ is a subset that is closed under vector addition and scalar multiplication.

1. Spanning: Every vector in $W$ can be written as a linear combination of the basis vectors. In notation, $\text{Span}{\mathbf{v}_1, \dots, \mathbf{v}_k} = W$.
2. Linear Independence: The only solution to $c_1\mathbf{v}_1 + \dots + c_k\mathbf{v}_k = \mathbf{0}$ is the trivial solution $c_1 = \dots = c_k = 0$.

The number of vectors in a basis is the dimension of the subspace, denoted $\dim(W)$. On the flip side, a subspace can have infinitely many different bases, but they all share the same number of vectors. Finding a basis is usually a procedural task; finding the standard basis or an orthogonal basis requires additional steps like the Gram-Schmidt process.

Scenario 1: Basis for the Column Space (Col A)

One of the most common tasks is finding a basis for the column space of a matrix $A$. Still, the column space is the span of the columns of $A$, but the columns themselves are rarely linearly independent. The standard algorithm relies on row reduction No workaround needed..

The Algorithm:

1. Row reduce matrix $A$ to its reduced row echelon form (RREF), denoted $R$.
2. Identify the pivot columns in $R$. These columns contain the leading 1s.
3. The corresponding columns in the original matrix $A$ form a basis for $\text{Col } A$.

Why the original matrix? Row operations change the column space. While the linear dependencies among columns are preserved, the actual vectors in the column space are not. The pivot columns of $A$ retain the exact linear relationships found in $R$ but exist in the correct vector space.

Example: Let $A = \begin{bmatrix} 1 & 2 & 3 \ 2 & 4 & 6 \ 1 & 2 & 4 \end{bmatrix}$. Row reducing yields $R = \begin{bmatrix} 1 & 2 & 0 \ 0 & 0 & 1 \ 0 & 0 & 0 \end{bmatrix}$. Pivot columns in $R$ are columns 1 and 3. So, a basis for $\text{Col } A$ is $\left{ \begin{bmatrix} 1 \ 2 \ 1 \end{bmatrix}, \begin{bmatrix} 3 \ 6 \ 4 \end{bmatrix} \right}$. The dimension (rank) is 2 Small thing, real impact..

Scenario 2: Basis for the Null Space (Nul A)

The null space of $A$ is the set of all solutions to the homogeneous equation $A\mathbf{x} = \mathbf{0}$. Since this is a solution set, it is automatically a subspace. Finding a basis here involves expressing the solution set in parametric vector form.

The Algorithm:

1. Row reduce $A$ to RREF.
2. Identify pivot variables (basic variables) and free variables.
3. Write the basic variables in terms of the free variables.
4. Express the solution vector $\mathbf{x}$ as a linear combination of vectors, where the coefficients are the free variables.
5. The vectors in this linear combination form a basis for $\text{Nul } A$.

Example: Using the RREF from above: $R = \begin{bmatrix} 1 & 2 & 0 \ 0 & 0 & 1 \ 0 & 0 & 0 \end{bmatrix}$. Equations: $x_1 + 2x_2 = 0 \Rightarrow x_1 = -2x_2$ and $x_3 = 0$. $x_2$ is free. $\mathbf{x} = \begin{bmatrix} -2x_2 \ x_2 \ 0 \end{bmatrix} = x_2 \begin{bmatrix} -2 \ 1 \ 0 \end{bmatrix}$. A basis for $\text{Nul } A$ is $\left{ \begin{bmatrix} -2 \ 1 \ 0 \end{bmatrix} \right}$. The dimension (nullity) is 1. Note the Rank-Nullity Theorem in action: $\text{rank}(A) + \text{nullity}(A) = 2 + 1 = 3 = n$ (number of columns) Less friction, more output..

Scenario 3: Basis for a Subspace Defined by Spanning Vectors

Often, a subspace $W$ is given explicitly as $W = \text{Span}{\mathbf{v}_1, \mathbf{v}_2, \dots, \mathbf{v}_p}$. The vectors are given, but they may be linearly dependent. To find a basis, we must discard the dependent vectors.

Method A: Row Space Technique (Standard)

1. Construct a matrix $A$ whose rows are the given vectors $\mathbf{v}_1, \dots, \mathbf{v}_p$.
2. Row reduce $A$ to RREF (or just Row Echelon Form).
3. The non-zero rows of the resulting matrix form a basis for the row space, which is exactly the span of the original vectors.

Method B: Column Space Technique (Alternative)

1. Construct a matrix $A$ whose columns are the given vectors.
2. Row reduce to RREF.
3. Identify pivot columns in the RREF.
4. Select the corresponding columns from the original matrix $A$.

Both methods work. Method A is often faster computationally because you stop at Row Echelon Form (REF) and simply read off the non-zero rows. Method B preserves the original vectors exactly as given, which is sometimes preferred for interpretation Still holds up..

Scenario 4: Basis for a Subspace Defined by Linear Equations

Sometimes a subspace is defined implicitly by a system of linear equations, such as: $W = { (x, y, z) \in \mathbb{R}^3 \mid x - 2y + 3z = 0 }$. This is geometrically a plane through the origin. Algebraically, it is the null space of the coefficient matrix.

The Algorithm:

1. Write the system as a homogeneous matrix equation $A\mathbf{x} = \mathbf{0}$.
2. Find a basis for $\text{Nul } A$ using the parametric vector form method described in Scenario 2.

Example: $W = { (x, y, z) \mid x - 2y + 3z = 0 }$. Matrix $A = \begin{bmatrix} 1 & -2 & 3 \end{bmatrix}$. Equation: $x = 2y - 3z$. $y, z$ are free. $\mathbf{x} = \begin{bmatrix} 2y - 3z \ y \ z \end{bmatrix} = y \begin{bmatrix} 2 \ 1 \ 0 \end{bmatrix} + z \begin{bmatrix} -3 \ 0 \ 1

\end{bmatrix}. A basis for $W$ is therefore [ \left{ \begin{bmatrix} 2 \ 1 \ 0 \end{bmatrix}, \begin{bmatrix} -3 \ 0 \ 1 \end{bmatrix} \right}. ] The dimension of $W$ is 2, which aligns with the geometric intuition that a plane through the origin in $\mathbb{R}^3$ is two-dimensional.

Summary and Conclusion

Finding a basis depends on how the subspace is presented:

• Null space: Solve $A\mathbf{x}=\mathbf{0}$, identify the free variables, and write the solution in parametric vector form. The vectors attached to the free variables form a basis for $\text{Nul } A$.
• Column space: Row reduce $A$ to RREF, identify the pivot columns, and select the corresponding columns from the original matrix $A$.
• Row space / spanning set: Place the given vectors as rows of a matrix, row reduce to REF or RREF, and take the non-zero rows as a basis.
• Subspace defined by equations: Rewrite the conditions as a homogeneous system $A\mathbf{x}=\mathbf{0}$ and apply the null space method.

In every case, a valid basis must be linearly independent and must span the subspace. The Rank-Nullity Theorem offers a crucial consistency check: $\text{rank}(A)+\text{nullity}(A)=n$, where $n$ is the number of columns of $A$. By mastering these techniques, one can translate fluidly between geometric descriptions, systems of linear equations, and matrix representations of subspaces.