Solved Exercises on Change of Basis Introduction

Master change of basis: coordinate systems, basis transformation, and coordinate conversion through these 5 detailed exercises.

Solution: Exercises 1 to 3
1 Standard to Custom Basis
Exercise 1
Let B = {b₁, b₂} be a basis where b₁ = (2, 1) and b₂ = (1, 1). Find the coordinate vector [v]_B of v = (5, 3) with respect to basis B.
Definition:

Coordinate vector: [v]_B = \(\begin{bmatrix} c₁ \\ c₂ \end{bmatrix}\) where v = c₁b₁ + c₂b₂

Method:
  1. Set up the equation v = c₁b₁ + c₂b₂
  2. Express as a system of linear equations
  3. Solve for the coefficients c₁, c₂
  4. Form the coordinate vector [v]_B
Set up equation
v = c₁b₁ + c₂b₂
Solve system
[v]_B = \(\begin{bmatrix} 2 \\ 1 \end{bmatrix}\)
Step 1: Set up the linear combination

We want to find c₁ and c₂ such that: v = c₁b₁ + c₂b₂

(5, 3) = c₁(2, 1) + c₂(1, 1)

(5, 3) = (2c₁ + c₂, c₁ + c₂)

Step 2: Form the system of equations

This gives us the system:

2c₁ + c₂ = 5 (equation 1)

c₁ + c₂ = 3 (equation 2)

Step 3: Solve the system

Subtract equation 2 from equation 1:

(2c₁ + c₂) - (c₁ + c₂) = 5 - 3

c₁ = 2

Substitute back into equation 2: 2 + c₂ = 3, so c₂ = 1

Step 4: Verify the solution

Check: 2b₁ + 1b₂ = 2(2, 1) + 1(1, 1) = (4, 2) + (1, 1) = (5, 3) ✓

Step 5: Form the coordinate vector

[v]_B = \(\begin{bmatrix} c₁ \\ c₂ \end{bmatrix} = \begin{bmatrix} 2 \\ 1 \end{bmatrix}\)

[v]_B = \(\begin{bmatrix} 2 \\ 1 \end{bmatrix}\)
Final answer:

The coordinate vector of v with respect to basis B is [v]_B = \(\begin{bmatrix} 2 \\ 1 \end{bmatrix}\)

Applied rules:

• Coordinate vector definition: v = Σ cᵢbᵢ

• Linear combination: Express vector as sum of basis vectors

• System solving: Solve for coefficients using elimination

2 Change of Basis Matrix
Exercise 2
Find the change of basis matrix from basis B = {(1, 2), (3, 4)} to basis C = {(1, 0), (0, 1)} (standard basis). Then convert [v]_B = \(\begin{bmatrix} 2 \\ 1 \end{bmatrix}\) to [v]_C.
Definition:

Change of basis matrix: P_{C←B} such that [v]_C = P_{C←B}[v]_B

Basis matrix
P = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\)
Convert coordinates
[v]_C = P[v]_B = \(\begin{bmatrix} 5 \\ 8 \end{bmatrix}\)
Step 1: Form the change of basis matrix

The change of basis matrix from B to C is formed by expressing each vector in B as a linear combination of vectors in C

Since C is the standard basis, the change of basis matrix P_{C←B} is simply the matrix whose columns are the vectors of B

P_{C←B} = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\) (columns are b₁ and b₂)

Step 2: Apply the change of basis formula

[v]_C = P_{C←B}[v]_B

[v]_C = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix} \begin{bmatrix} 2 \\ 1 \end{bmatrix}\)

Step 3: Perform the matrix multiplication

Row 1: (1)(2) + (3)(1) = 2 + 3 = 5

Row 2: (2)(2) + (4)(1) = 4 + 4 = 8

So [v]_C = \(\begin{bmatrix} 5 \\ 8 \end{bmatrix}\)

Step 4: Verify by converting back to vector form

v = 5(1, 0) + 8(0, 1) = (5, 0) + (0, 8) = (5, 8)

Check using B coordinates: v = 2(1, 2) + 1(3, 4) = (2, 4) + (3, 4) = (5, 8) ✓

Step 5: Generalize the process

For any basis B = {b₁, b₂, ..., bₙ}, the change of basis matrix to standard basis has bᵢ as column i

This matrix converts B-coordinates to standard coordinates

P_{C←B} = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\), [v]_C = \(\begin{bmatrix} 5 \\ 8 \end{bmatrix}\)
Final answer:

The change of basis matrix is P_{C←B} = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\) and [v]_C = \(\begin{bmatrix} 5 \\ 8 \end{bmatrix}\)

Applied rules:

• Change of basis formula: [v]_C = P_{C←B}[v]_B

• Standard basis: Change matrix has basis vectors as columns

• Coordinate conversion: Matrix multiplication transforms coordinates

3 Inverse Change of Basis
Exercise 3
Given the same bases B = {(1, 2), (3, 4)} and C = {(1, 0), (0, 1)}, find the change of basis matrix from C to B, i.e., P_{B←C}. Then convert [w]_C = \(\begin{bmatrix} 7 \\ 10 \end{bmatrix}\) to [w]_B.
Definition:

Inverse change of basis: P_{B←C} = (P_{C←B})⁻¹

P_{C←B}
P_{C←B} = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\)
P_{B←C}
P_{B←C} = \(\begin{bmatrix} -2 & 3/2 \\ 1 & -1/2 \end{bmatrix}\)
Step 1: Find P_{B←C} = (P_{C←B})⁻¹

From Exercise 2: P_{C←B} = \(\begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix}\)

det(P_{C←B}) = (1)(4) - (3)(2) = 4 - 6 = -2

Using the 2×2 inverse formula: A⁻¹ = (1/det(A))\(\begin{bmatrix} d & -b \\ -c & a \end{bmatrix}\)

P_{B←C} = \(\frac{1}{-2}\begin{bmatrix} 4 & -3 \\ -2 & 1 \end{bmatrix} = \begin{bmatrix} -2 & 3/2 \\ 1 & -1/2 \end{bmatrix}\)

Step 2: Apply the change of basis formula

[w]_B = P_{B←C}[w]_C

[w]_B = \(\begin{bmatrix} -2 & 3/2 \\ 1 & -1/2 \end{bmatrix} \begin{bmatrix} 7 \\ 10 \end{bmatrix}\)

Step 3: Perform the matrix multiplication

Row 1: (-2)(7) + (3/2)(10) = -14 + 15 = 1

Row 2: (1)(7) + (-1/2)(10) = 7 - 5 = 2

So [w]_B = \(\begin{bmatrix} 1 \\ 2 \end{bmatrix}\)

Step 4: Verify the result

Convert back to check: w = 1(1, 2) + 2(3, 4) = (1, 2) + (6, 8) = (7, 10) ✓

This matches [w]_C = \(\begin{bmatrix} 7 \\ 10 \end{bmatrix}\) in standard coordinates

Step 5: Verify the inverse relationship

P_{B←C} × P_{C←B} should equal I:

\(\begin{bmatrix} -2 & 3/2 \\ 1 & -1/2 \end{bmatrix} \begin{bmatrix} 1 & 3 \\ 2 & 4 \end{bmatrix} = \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix} = I\) ✓

[w]_B = \(\begin{bmatrix} 1 \\ 2 \end{bmatrix}\)
Final answer:

The coordinate vector [w]_B = \(\begin{bmatrix} 1 \\ 2 \end{bmatrix}\)

Applied rules:

• Inverse relationship: P_{B←C} = (P_{C←B})⁻¹

• Verification: P_{B←C}P_{C←B} = I

• Coordinate conversion: [v]_B = P_{B←C}[v]_C

Change of Basis: Key Concepts
[v]_C = P_{C←B}[v]_B
Change of Basis Formula
Formula
P_{C←B} = C^{-1}B
General Change Matrix
Property
P_{B←C} = (P_{C←B})^{-1}
Inverse Property
Property
P_{C←B}P_{B←C} = I
Inverse Product
Coordinate invariance: The vector itself doesn't change, only its representation
Basis independence: Vector spaces exist independently of coordinate systems
Linear transformation: Change of basis is a linear transformation on coordinate vectors
Solution: Exercises 4 to 5
4 Transition Between Two Non-Standard Bases
Exercise 4
Let B = {(1, 1), (1, -1)} and D = {(2, 1), (1, 2)}. Find the change of basis matrix from B to D. Then convert [v]_B = \(\begin{bmatrix} 3 \\ 1 \end{bmatrix}\) to [v]_D.
Definition:

Transition matrix: P_{D←B} such that [v]_D = P_{D←B}[v]_B

P_{D←B}
P_{D←B} = D^{-1}B
Calculate
[v]_D = \(\begin{bmatrix} 2 \\ 1 \end{bmatrix}\)
Step 1: Form matrices B and D

B = \(\begin{bmatrix} 1 & 1 \\ 1 & -1 \end{bmatrix}\) (columns are vectors in basis B)

D = \(\begin{bmatrix} 2 & 1 \\ 1 & 2 \end{bmatrix}\) (columns are vectors in basis D)

Step 2: Calculate D⁻¹

For D = \(\begin{bmatrix} 2 & 1 \\ 1 & 2 \end{bmatrix}\): det(D) = (2)(2) - (1)(1) = 4 - 1 = 3

D⁻¹ = \(\frac{1}{3}\begin{bmatrix} 2 & -1 \\ -1 & 2 \end{bmatrix} = \begin{bmatrix} 2/3 & -1/3 \\ -1/3 & 2/3 \end{bmatrix}\)

Step 3: Calculate P_{D←B} = D⁻¹B

P_{D←B} = \(\begin{bmatrix} 2/3 & -1/3 \\ -1/3 & 2/3 \end{bmatrix} \begin{bmatrix} 1 & 1 \\ 1 & -1 \end{bmatrix}\)

Row 1, Col 1: (2/3)(1) + ((-1/3)(1) = 2/3 - 1/3 = 1/3

Row 1, Col 2: (2/3)(1) + ((-1/3)(-1) = 2/3 + 1/3 = 1

Row 2, Col 1: ((-1/3)(1) + (2/3)(1) = -1/3 + 2/3 = 1/3

Row 2, Col 2: ((-1/3)(1) + (2/3)(-1) = -1/3 - 2/3 = -1

So P_{D←B} = \(\begin{bmatrix} 1/3 & 1 \\ 1/3 & -1 \end{bmatrix}\)

Step 4: Convert coordinates

[v]_D = P_{D←B}[v]_B = \(\begin{bmatrix} 1/3 & 1 \\ 1/3 & -1 \end{bmatrix} \begin{bmatrix} 3 \\ 1 \end{bmatrix}\)

Row 1: (1/3)(3) + (1)(1) = 1 + 1 = 2

Row 2: (1/3)(3) + (-1)(1) = 1 - 1 = 0

So [v]_D = \(\begin{bmatrix} 2 \\ 0 \end{bmatrix}\)

Step 5: Verify the result

Check in original coordinates:

From B: v = 3(1, 1) + 1(1, -1) = (3, 3) + (1, -1) = (4, 2)

From D: v = 2(2, 1) + 0(1, 2) = (4, 2) + (0, 0) = (4, 2) ✓

P_{D←B} = \(\begin{bmatrix} 1/3 & 1 \\ 1/3 & -1 \end{bmatrix}\), [v]_D = \(\begin{bmatrix} 2 \\ 0 \end{bmatrix}\)
Final answer:

The change of basis matrix is P_{D←B} = \(\begin{bmatrix} 1/3 & 1 \\ 1/3 & -1 \end{bmatrix}\) and [v]_D = \(\begin{bmatrix} 2 \\ 0 \end{bmatrix}\)

Applied rules:

• General formula: P_{D←B} = D⁻¹B

• Matrix multiplication: Standard algorithm

• Verification: Convert back to standard coordinates

5 Orthonormal Basis Conversion
Exercise 5
Let E = {(1, 0), (0, 1)} be the standard basis and F = {(\(\frac{1}{\sqrt{2}}, \frac{1}{\sqrt{2}}\)), (\(-\frac{1}{\sqrt{2}}, \frac{1}{\sqrt{2}}\))} be a rotation of E by 45°. Find [v]_F for v = (3, 4) in standard coordinates.
Definition:

Orthonormal basis: Basis vectors are orthogonal and have unit length

F matrix
F = \(\begin{bmatrix} 1/\sqrt{2} & -1/\sqrt{2} \\ 1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix}\)
Since F is orthogonal
F⁻¹ = Fᵀ
Step 1: Form the basis matrix F

F = \(\begin{bmatrix} 1/\sqrt{2} & -1/\sqrt{2} \\ 1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix}\)

This is a rotation matrix by 45°, which is orthogonal

Step 2: Verify F is orthogonal

Check if FᵀF = I:

Fᵀ = \(\begin{bmatrix} 1/\sqrt{2} & 1/\sqrt{2} \\ -1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix}\)

FᵀF = \(\begin{bmatrix} 1/\sqrt{2} & 1/\sqrt{2} \\ -1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix} \begin{bmatrix} 1/\sqrt{2} & -1/\sqrt{2} \\ 1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix}\)

= \(\begin{bmatrix} 1/2+1/2 & -1/2+1/2 \\ -1/2+1/2 & 1/2+1/2 \end{bmatrix} = \begin{bmatrix} 1 & 0 \\ 0 & 1 \end{bmatrix} = I\) ✓

Step 3: Use the property of orthogonal matrices

Since F is orthogonal: F⁻¹ = Fᵀ

So P_{F←E} = F⁻¹ = Fᵀ = \(\begin{bmatrix} 1/\sqrt{2} & 1/\sqrt{2} \\ -1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix}\)

Step 4: Calculate [v]_F = Fᵀv

[v]_F = \(\begin{bmatrix} 1/\sqrt{2} & 1/\sqrt{2} \\ -1/\sqrt{2} & 1/\sqrt{2} \end{bmatrix} \begin{bmatrix} 3 \\ 4 \end{bmatrix}\)

Row 1: (1/√2)(3) + (1/√2)(4) = (3 + 4)/√2 = 7/√2

Row 2: (-1/√2)(3) + (1/√2)(4) = (-3 + 4)/√2 = 1/√2

So [v]_F = \(\begin{bmatrix} 7/\sqrt{2} \\ 1/\sqrt{2} \end{bmatrix}\)

Step 5: Verify the result

Convert back: v = (7/√2)f₁ + (1/√2)f₂

= (7/√2)(1/√2, 1/√2) + (1/√2)(-1/√2, 1/√2)

= (7/2, 7/2) + (-1/2, 1/2) = (3, 4) ✓

[v]_F = \(\begin{bmatrix} 7/\sqrt{2} \\ 1/\sqrt{2} \end{bmatrix}\)
Final answer:

[v]_F = \(\begin{bmatrix} 7/\sqrt{2} \\ 1/\sqrt{2} \end{bmatrix}\)

Applied rules:

• Orthogonal matrices: F⁻¹ = Fᵀ

• Rotation matrices: Preserve length and angles

• Orthonormal coordinates: Simplified transformation formulas

Change of Basis: Comprehensive Guide
[v]_{B'} = P_{B' \leftarrow B}[v]_B
General Change Formula
Key definitions:

Basis: Linearly independent set that spans the vector space

Coordinate vector: Representation of a vector with respect to a basis

Change of basis matrix: Matrix that converts coordinates from one basis to another

Complete methodology:
  1. Identify the bases: Specify B and B' clearly
  2. Form basis matrices: Create matrices with basis vectors as columns
  3. Calculate change matrix: P_{B'←B} = (B')⁻¹B
  4. Apply transformation: [v]_{B'} = P_{B'←B}[v]_B
  5. Verify result: Convert back to check
Tip 1: Always check that your basis vectors are linearly independent.
Tip 2: For orthogonal bases, use the transpose property: A⁻¹ = Aᵀ.
Tip 3: Remember: P_{B←B'} = (P_{B'←B})⁻¹.
Tip 4: The identity matrix converts coordinates to standard basis.
Common errors: Confusing the direction of transformation, miscalculating inverses, forgetting to verify.
Exam preparation: Master 2×2 cases, understand geometric interpretation.
Properties to know by heart:

• Coordinate invariance: The vector remains the same object

• Inverse property: P_{B←B'} = (P_{B'←B})⁻¹

• Chain rule: P_{C←A} = P_{C←B}P_{B←A}

• Identity: P_{B←B} = I

• Orthogonal basis: P⁻¹ = Pᵀ

Basis Change Visualization
Exercise 6: Coordinate Transformation Analysis
Visual analysis of coordinate transformations:
• Different basis systems
• Coordinate vector relationships
• Transformation properties

Analysis: The visualization shows how different basis systems affect coordinate representations.

  • Standard basis provides familiar coordinates
  • Custom bases can simplify problems
  • Orthogonal bases preserve geometric relationships

Questions & Answers

Question: Why do we need change of basis? Isn't the standard basis good enough for everything?

Answer: While the standard basis is convenient, other bases are often more useful:

Problem-specific advantages:

  • Eigenvector basis: Makes matrix diagonal, simplifying powers and exponentials
  • Fourier basis: Ideal for signal processing and periodic phenomena
  • Principal components: Reveals the most important directions in data
  • Wavelet basis: Useful for image compression and analysis

Geometric insights:

Some problems have natural coordinate systems. For example, planetary orbits are better described in polar coordinates than Cartesian coordinates.

Computational efficiency:

Changing to an appropriate basis can dramatically simplify calculations. A complex transformation matrix in standard coordinates might become diagonal in the right basis.

Change of basis is like choosing the right language to describe a phenomenon - it doesn't change the underlying reality, but it can make the description much clearer.

Question: How do I remember which matrix goes where in P_{C←B}? It's confusing.

Answer: Use the mnemonic "From Right to Left":

Notation: P_{Target←Source}

This means: P_{C←B} converts coordinates FROM basis B TO basis C

Formula: [v]_C = P_{C←B}[v]_B

The subscript tells you: [input]_B → [output]_C

Alternative memory aid:

Think of it as "B-to-C matrix" for converting B-coordinates to C-coordinates

Calculation:

P_{C←B} = C⁻¹B, where C and B are matrices with basis vectors as columns

This formula makes sense: we want to express B-vectors in terms of C-vectors, so we multiply by C⁻¹.

Always remember: the matrix goes on the left of the input coordinate vector, and the output has the subscript that matches the "target" in the notation.

Question: What are practical applications of change of basis?

Answer: Change of basis has numerous practical applications:

  • Computer Graphics: Converting between world coordinates, camera coordinates, and screen coordinates
  • Quantum Mechanics: Switching between position and momentum representations
  • Data Science: Principal Component Analysis (PCA) for dimensionality reduction
  • Signal Processing: Fourier transforms and wavelet transforms
  • Robotics: Converting between joint space and Cartesian space
  • Machine Learning: Feature transformation and dimensionality reduction

Specific examples:

In image compression, we change to a basis where most information is concentrated in few coefficients (JPEG uses discrete cosine transform).

In quantum mechanics, the Schrödinger equation might be easier to solve in momentum space than position space.

In structural engineering, we find the principal axes of stress to understand material failure.

The key insight is that while the mathematical object (vector, transformation) remains the same, choosing the right coordinate system can make problems tractable that would otherwise be intractable.