Linear Algebra: Coordinate Systems and Representations

Vectors are more than just arrows or lists of numbers; they are abstract objects that live in spaces defined by algebraic rules. To perform calculations with these objects—whether you're analyzing forces in a bridge, processing signals, or training a machine learning model—you need a concrete, numerical representation. This is where coordinate systems and basis choices come in, providing the crucial link between abstract vector theory and the computational power of $R^{n}$ .

1. The Bridge to Computation: Coordinates Relative to a Basis

A basis $B = {v_{1}, v_{2}, ..., v_{n}}$ for a vector space $V$ is a set of linearly independent vectors that span $V$ . Every vector $x$ in $V$ can be expressed uniquely as a linear combination of the basis vectors: $x = c_{1} v_{1} + c_{2} v_{2} + ... + c_{n} v_{n} .$ The scalars $c_{1}, c_{2}, ..., c_{n}$ are called the coordinates of $x$ relative to the basis $B$ , and the column vector $[x]_{B} = c_{1} c_{2} ⋮ c_{n}$ is the coordinate vector of $x$ .

Think of the basis as a set of reference directions or a custom ruler. The standard basis in $R^{2}$ , ${\hat{i}, \hat{j}}$ , gives you the familiar $(x, y)$ coordinates. However, in a different basis, the same geometric point has a different coordinate vector. For example, in $R^{2}$ , the vector $x = (5, 5)$ in the standard basis has coordinates $[x]_{B} = [11]$ relative to the basis $B = {(5, 0), (0, 5)}$ , because $(5, 5) = 1 * (5, 0) + 1 * (0, 5)$ .

2. The Coordinate Mapping: An Isomorphism with $R^{n}$

The process of taking an abstract vector $x$ and producing its coordinate vector $[x]_{B}$ is a function called the coordinate mapping (or coordinate transformation). This mapping, denoted $x \mapsto [x]_{B}$ , is a one-to-one and onto linear transformation from the vector space $V$ to $R^{n}$ .

This relationship is formally an isomorphism—a structure-preserving bijection. If two vector spaces are isomorphic (like any real $n$ -dimensional space $V$ and $R^{n}$ ), they are algebraically identical. Every vector addition or scalar multiplication in $V$ corresponds perfectly to the same operation on the coordinate vectors in $R^{n}$ . This is the foundational theorem that allows engineers and scientists to replace abstract vector manipulations with concrete matrix arithmetic on lists of numbers. It confirms that $R^{n}$ is the universal computational model for all finite-dimensional real vector spaces.

3. The Computational Workhorse: Finding Coordinates via Row Reduction

How do you actually find the coordinate vector $[x]_{B}$ ? The most systematic method sets up and solves a linear system via row reduction. Given a basis $B = {b_{1}, ..., b_{n}}$ for $V$ and a vector $x$ in $V$ , you want scalars $c_{1}, ..., c_{n}$ such that: $c_{1} b_{1} + ... + c_{n} b_{n} = x .$

In $R^{n}$ , you can form a matrix $P_{B}$ whose columns are the basis vectors $b_{1}, ..., b_{n}$ . This matrix is called the change-of-coordinates matrix (from $B$ to the standard basis). The equation becomes: $P_{B} [x]_{B} = x .$ To solve for $[x]_{B}$ , you form an augmented matrix $[P_{B} ∣ x]$ and row reduce it to reduced echelon form.

Example: Let $B = {[11], [1 - 1]}$ and $x = [31]$ . Find $[x]_{B}$ .

Set up the augmented system:

$[11 1 - 1 31]$

Row reduce: $R 2 \leftarrow R 2 - R 1$ yields $[10 1 - 2 3 - 2]$ .

Then $R 2 \leftarrow - \frac{1}{2} R 2$ : $[101131]$ . Finally, $R 1 \leftarrow R 1 - R 2$ : $[100121]$ .

The solution is $c_{1} = 2, c_{2} = 1$ . Therefore, $[x]_{B} = [21]$ .

Verification: $2 [11] + 1 [1 - 1] = [31]$ .

4. Basis Choice in Applications: Simplifying Complexity

The choice of basis is not merely a mathematical formality; it is a powerful tool for simplifying problems. A well-chosen basis can diagonalize a matrix, decouple differential equations, or compress data with minimal loss.

Principal Component Analysis (PCA): In data science, PCA finds a new orthonormal basis (the principal components) for a dataset. The first basis vector points in the direction of maximum variance in the data. Representing data in this new coordinate system often allows you to discard coordinates (dimensions) with negligible variance, effectively reducing dimensionality and noise.
Fourier Analysis: Representing a signal (like an audio wave) in the standard time basis shows its amplitude over time. Representing the same signal in a Fourier basis (composed of sine and cosine functions of different frequencies) shows its frequency spectrum. This coordinate change transforms convoluted differential operations in the time domain into simple multiplications in the frequency domain.
Mechanics and Eigenvectors: In structural engineering, analyzing vibrations often involves finding the eigenvectors of a stiffness matrix. These eigenvectors form a natural basis for the system. Representing the system's state in this eigenbasis diagonalizes the equations of motion, revealing independent normal modes of vibration—each with its own characteristic frequency.

The coordinates of a vector are meaningless without specifying the basis. The power of linear algebra lies in strategically changing bases to make the problem at hand as simple as possible.

Common Pitfalls

Confusing the Vector with its Coordinate Vector: A vector $v$ is an abstract object. Its coordinate vector $[v]_{B}$ is a representation of that object relative to a specific basis $B$ . They are not the same thing, just as a person is not the same as their street address. Always be clear about which basis you are using.

Correction: Explicitly write the basis subscript, e.g., $[v]_{B}$ vs. $[v]_{C}$ . Never refer to a coordinate vector as "the vector" without context.

Assuming the Standard Basis: A common error is to assume coordinates are always given relative to the standard basis ${\hat{i}, \hat{j}, \hat{k}}$ . In many applications, like computer graphics or quantum mechanics, you work exclusively in non-standard bases.

Correction: When presented with a vector as a column of numbers, always ask: "With respect to what basis?" If unspecified in a problem, the standard basis is usually implied, but this should be a conscious check.

Incorrectly Ordering the Basis: The coordinate vector $[2, 1]^{T}$ means the first basis vector is multiplied by 2 and the second by 1. If you list the basis set in a different order, the coordinate vector for the same abstract vector will be completely different.

Correction: When defining a basis $B$ , treat it as an ordered list, not just a set. Maintain this order consistently throughout all calculations.

Misapplying Row Reduction: When setting up the augmented matrix $[P_{B} ∣ x]$ , the columns of $P_{B}$ must be the basis vectors $b_{i}$ . A frequent mistake is to place them as rows, which solves a different, unrelated system.

Correction: Remember the matrix equation: $P_{B} [x]_{B} = x$ . The unknown $[x]_{B}$ is a vector of weights that linearly combines the columns of $P_{B}$ .

Summary

The coordinate vector $[x]_{B}$ provides a numerical representation of an abstract vector $x$ relative to an ordered basis $B$ .
The coordinate mapping $x \mapsto [x]_{B}$ is an isomorphism, proving that any finite-dimensional real vector space can be computed with using $R^{n}$ .
The primary method for finding coordinates is solving the linear system $P_{B} [x]_{B} = x$ via row reduction on the augmented matrix $[P_{B} ∣ x]$ .
The choice of basis is a powerful application-driven decision. Transformations like PCA and the Fourier transform are essentially changes of basis that simplify analysis by revealing a problem's underlying structure.
Always specify your basis. A vector's coordinates are entirely dependent on this choice, and confusing a vector with its coordinate representation is a fundamental error.

Linear Algebra: Coordinate Systems and Representations

1. The Bridge to Computation: Coordinates Relative to a Basis

A basis $B = \{\vec{v}_1, \vec{v}_2, ..., \vec{v}_n\}$ for a vector space $V$ is a set of linearly independent vectors that span $V$ . Every vector $\vec{x}$ in $V$ can be expressed uniquely as a linear combination of the basis vectors: $$\vec{x} = c_1\vec{v}_1 + c_2\vec{v}_2 + ... + c_n\vec{v}_n.$$ The scalars $c_1, c_2, ..., c_n$ are called the coordinates of $\vec{x}$ relative to the basis $B$ , and the column vector $$[\vec{x}]_B = \begin{bmatrix} c_1 \\ c_2 \\ \vdots \\ c_n \end{bmatrix}$$ is the coordinate vector of $\vec{x}$ .

Think of the basis as a set of reference directions or a custom ruler. The standard basis in $R^2$ , $\{\hat{i}, \hat{j}\}$ , gives you the familiar $(x, y)$ coordinates. However, in a different basis, the same geometric point has a different coordinate vector. For example, in $R^2$ , the vector $\vec{x} = (5, 5)$ in the standard basis has coordinates $[\vec{x}]_B = \begin{bmatrix} 1 \\ 1 \end{bmatrix}$ relative to the basis $B = \{(5,0), (0,5)\}$ , because $(5,5) = 1*(5,0) + 1*(0,5)$ .

2. The Coordinate Mapping: An Isomorphism with $R^n$

The process of taking an abstract vector $\vec{x}$ and producing its coordinate vector $[\vec{x}]_B$ is a function called the coordinate mapping (or coordinate transformation). This mapping, denoted $\vec{x} \mapsto [\vec{x}]_B$ , is a one-to-one and onto linear transformation from the vector space $V$ to $R^n$ .

This relationship is formally an isomorphism—a structure-preserving bijection. If two vector spaces are isomorphic (like any real $n$ -dimensional space $V$ and $R^n$ ), they are algebraically identical. Every vector addition or scalar multiplication in $V$ corresponds perfectly to the same operation on the coordinate vectors in $R^n$ . This is the foundational theorem that allows engineers and scientists to replace abstract vector manipulations with concrete matrix arithmetic on lists of numbers. It confirms that $R^n$ is the universal computational model for all finite-dimensional real vector spaces.

3. The Computational Workhorse: Finding Coordinates via Row Reduction

How do you actually find the coordinate vector $[\vec{x}]_B$ ? The most systematic method sets up and solves a linear system via row reduction. Given a basis $B = \{\vec{b}_1, ..., \vec{b}_n\}$ for $V$ and a vector $\vec{x}$ in $V$ , you want scalars $c_1, ..., c_n$ such that: $$c_1\vec{b}_1 + ... + c_n\vec{b}_n = \vec{x}.$$

In $R^n$ , you can form a matrix $P_B$ whose columns are the basis vectors $\vec{b}_1, ..., \vec{b}_n$ . This matrix is called the change-of-coordinates matrix (from $B$ to the standard basis). The equation becomes: $$P_B [\vec{x}]_B = \vec{x}.$$ To solve for $[\vec{x}]_B$ , you form an augmented matrix $[P_B | \vec{x}]$ and row reduce it to reduced echelon form.

Example: Let $B = \left\{ \begin{bmatrix} 1 \\ 1 \end{bmatrix}, \begin{bmatrix} 1 \\ -1 \end{bmatrix} \right\}$ and $\vec{x} = \begin{bmatrix} 3 \\ 1 \end{bmatrix}$ . Find $[\vec{x}]_B$ .

Set up the augmented system:

$$\left[ \begin{array}{cc|c} 1 & 1 & 3 \\ 1 & -1 & 1 \end{array} \right]$$

Row reduce: $R2 \leftarrow R2 - R1$ yields $\left[ \begin{array}{cc|c} 1 & 1 & 3 \\ 0 & -2 & -2 \end{array} \right]$ .

Then $R2 \leftarrow -\frac{1}{2}R2$ : $\left[ \begin{array}{cc|c} 1 & 1 & 3 \\ 0 & 1 & 1 \end{array} \right]$ . Finally, $R1 \leftarrow R1 - R2$ : $\left[ \begin{array}{cc|c} 1 & 0 & 2 \\ 0 & 1 & 1 \end{array} \right]$ .

The solution is $c_1=2, c_2=1$ . Therefore, $[\vec{x}]_B = \begin{bmatrix} 2 \\ 1 \end{bmatrix}$ .

Verification: $2\begin{bmatrix}1\\1\end{bmatrix} + 1\begin{bmatrix}1\\-1\end{bmatrix} = \begin{bmatrix}3\\1\end{bmatrix}$ .

4. Basis Choice in Applications: Simplifying Complexity

Principal Component Analysis (PCA): In data science, PCA finds a new orthonormal basis (the principal components) for a dataset. The first basis vector points in the direction of maximum variance in the data. Representing data in this new coordinate system often allows you to discard coordinates (dimensions) with negligible variance, effectively reducing dimensionality and noise.
Fourier Analysis: Representing a signal (like an audio wave) in the standard time basis shows its amplitude over time. Representing the same signal in a Fourier basis (composed of sine and cosine functions of different frequencies) shows its frequency spectrum. This coordinate change transforms convoluted differential operations in the time domain into simple multiplications in the frequency domain.
Mechanics and Eigenvectors: In structural engineering, analyzing vibrations often involves finding the eigenvectors of a stiffness matrix. These eigenvectors form a natural basis for the system. Representing the system's state in this eigenbasis diagonalizes the equations of motion, revealing independent normal modes of vibration—each with its own characteristic frequency.

The coordinates of a vector are meaningless without specifying the basis. The power of linear algebra lies in strategically changing bases to make the problem at hand as simple as possible.

Common Pitfalls

Confusing the Vector with its Coordinate Vector: A vector $\vec{v}$ is an abstract object. Its coordinate vector $[\vec{v}]_B$ is a representation of that object relative to a specific basis $B$ . They are not the same thing, just as a person is not the same as their street address. Always be clear about which basis you are using.

Correction: Explicitly write the basis subscript, e.g., $[\vec{v}]_B$ vs. $[\vec{v}]_C$ . Never refer to a coordinate vector as "the vector" without context.

Assuming the Standard Basis: A common error is to assume coordinates are always given relative to the standard basis $\{\hat{i}, \hat{j}, \hat{k}\}$ . In many applications, like computer graphics or quantum mechanics, you work exclusively in non-standard bases.

Correction: When presented with a vector as a column of numbers, always ask: "With respect to what basis?" If unspecified in a problem, the standard basis is usually implied, but this should be a conscious check.

Incorrectly Ordering the Basis: The coordinate vector $[2, 1]^T$ means the first basis vector is multiplied by 2 and the second by 1. If you list the basis set in a different order, the coordinate vector for the same abstract vector will be completely different.

Correction: When defining a basis $B$ , treat it as an ordered list, not just a set. Maintain this order consistently throughout all calculations.

Misapplying Row Reduction: When setting up the augmented matrix $[P_B | \vec{x}]$ , the columns of $P_B$ must be the basis vectors $\vec{b}_i$ . A frequent mistake is to place them as rows, which solves a different, unrelated system.

Correction: Remember the matrix equation: $P_B [\vec{x}]_B = \vec{x}$ . The unknown $[\vec{x}]_B$ is a vector of weights that linearly combines the columns of $P_B$ .

Summary

The coordinate vector $[\vec{x}]_B$ provides a numerical representation of an abstract vector $\vec{x}$ relative to an ordered basis $B$ .
The coordinate mapping $\vec{x} \mapsto [\vec{x}]_B$ is an isomorphism, proving that any finite-dimensional real vector space can be computed with using $R^n$ .
The primary method for finding coordinates is solving the linear system $P_B [\vec{x}]_B = \vec{x}$ via row reduction on the augmented matrix $[P_B | \vec{x}]$ .
The choice of basis is a powerful application-driven decision. Transformations like PCA and the Fourier transform are essentially changes of basis that simplify analysis by revealing a problem's underlying structure.
Always specify your basis. A vector's coordinates are entirely dependent on this choice, and confusing a vector with its coordinate representation is a fundamental error.

Linear Algebra: Coordinate Systems and Representations

Linear Algebra: Coordinate Systems and Representations

1. The Bridge to Computation: Coordinates Relative to a Basis

2. The Coordinate Mapping: An Isomorphism with Rn

3. The Computational Workhorse: Finding Coordinates via Row Reduction

4. Basis Choice in Applications: Simplifying Complexity

Common Pitfalls

Summary

Linear Algebra: Coordinate Systems and Representations

1. The Bridge to Computation: Coordinates Relative to a Basis

2. The Coordinate Mapping: An Isomorphism with $R^n$

3. The Computational Workhorse: Finding Coordinates via Row Reduction

4. Basis Choice in Applications: Simplifying Complexity

Common Pitfalls

Summary

Write better notes with AI

Linear Algebra: Coordinate Systems and Representations

1. The Bridge to Computation: Coordinates Relative to a Basis

2. The Coordinate Mapping: An Isomorphism with Rn

3. The Computational Workhorse: Finding Coordinates via Row Reduction

4. Basis Choice in Applications: Simplifying Complexity

Common Pitfalls

Summary

2. The Coordinate Mapping: An Isomorphism with $R^{n}$

2. The Coordinate Mapping: An Isomorphism with $R^{n}$