Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
78 Cards in this Set
- Front
- Back
If a transformation is through an n x m matrix, what is the dimensional change?
|
From R^m to R^n
|
|
When is k |x|
|x| a dilation and when is it a contraction? |
Dilation - when K is greater than 1
Contraction - when K is between 0 and 1 |
|
How can you describe a projection in words?
|
proj(x) means turning x into something x(parallel) and x(perpendicular)
2.1 |
|
How do you calculate a projection onto a line spanned by a vector?
|
1. Normalize the vector
2. proj(x) = (x•u)u |
|
How do you describe ref(x) algebraically?
|
ref(x) = 2proj(x) - x = 2(x•u)u - x
|
|
How do you describe a reflection in words?
|
Flipping a vector over something = x(parallel) - x(perpendicular)
|
|
What conditions have to be fulfilled in order for a matrix to be invertibel?
|
1. It must be square
2. rref(A) = 1 |
|
If T is transofmred from R^m to R^n, describe the image and the kernel.
|
im(T) is the subset of the codomain R^n
ker(T) is the domain of R^m |
|
How do you describe the inverse of a 2x2 matrix, A = |a b|
|c d| |
A^-1 = 1/(det(A)) |d -b|
|-c a| |
|
If it is in the form y1 = ax1 + bx2 + cx3, what form is the inverse in?
|
Inverse is in the form, x1 = ay1 + by2 + cx3
|
|
What is hte kernel of an invertible nxn matrix?
|
ker = {0}
|
|
Is im(A) necessarily equal to im(rref(A))?
|
No
|
|
For an invertible nxn matrix, is im(A) = R^n?
|
im(A) = R^n
|
|
What is the rank?
|
Dimension of the image.
|
|
Is the image dependent on kernel?
|
No
|
|
What does it mean to be orthogonal?
|
Dot product is 0
|
|
What does orthonormal mean?
|
Unit vector are orthogonal to one another.
|
|
Are orthonormal vectors linearly independent?
|
Yes
|
|
What do orthnormal vectors form?
|
Basis of R^n.
|
|
How would you describe proj(x) in words?
|
An orthogonal projection amounts to porjectings it onto the x-axis, then onto the y-axis, then adding the resultant vectors.
|
|
How does you describe proj(x) algebraically?
|
(u1 • x1) u1 + ... + (uN • xN) uN
|
|
What is orthogonal complement?
|
The set of vectors in R^n that are orthogonal to all vectors in V.
|
|
What is the orthogonal complement the kernel of?
|
The orthogonal projection on to V.
|
|
What is the dimension of the subspace + the dimsneion of the kernel equal to?
For example, what is the kernel of a plane in R^3? What is the kernel of a line in R^3? |
0
Line: 1 + 2 = 3 Plane: 2 + 1 = 3 (4.1) |
|
What is V ∩ V⊥?
|
0
|
|
What condition do 2 vectors x and y fit ≡ they are orthogonal complements?
|
||x + y||^2 = ||x||^2 + ||y||^2
|
|
Which is always greater than or equal to the other: x or proj(x)?
|
x by the Pythagorean Theorem
|
|
What is the angel between 2 vectors?
|
arc cos (x•y/(||x|| ||y||))
|
|
What is the correlation coefficient between two vectors?
|
r = cos(Θ) = (x•y/(||x|| ||y||)
(4.2) |
|
What is another way to write v1 • v2 using Θ?
|
v1•v2 = ||v1|| ||v2|| cos(Θ)
|
|
If basis of V is v1 and v2, then how would you find the orthonormal basis?
|
1. u1 = normalized v1
2. u2 = u2 - (u1•v2)u1 (v3 would equal u3 - (u1•v3)u1 - (u2•v3)u2/ ||"|| ) |
|
Describe the relationship between basis and image with the kernel.
|
Basis is found in the kernel, so it involves the kernel.
Image does not involve the kernel. |
|
What is an orthonormal transformation?
|
1. Preserves length
2. If put something with length one in, get length one out 3. All of its columns are linearly independent 4. All of its columns have length one. |
|
For an nxn matrix, how can you compare the determinants of A and A^T?
|
det(A) = det(A^T)
|
|
What do the columns of an orthogonal matrix form?
|
An orthonormal basis of R^n
|
|
Is the product of 2 orthogonal matrices also orthogonal?
|
Yes
|
|
Is the inverse of an orthogonal matrix orthogonal?
|
Yes
|
|
Is matrix A is nxm, what is a matrix A^T?
|
mxn
|
|
What is a square matrix?
|
A = A^T
|
|
What is a skew-symmetric matrix?
|
-A = A^T
|
|
What is the general form a of skew matrix?
|
0 a b
-a 0 c -b -c 0 |
|
If a matrix is orthogonal, what is A x A^T?
|
Identity Matrix
|
|
Is rank(A) = rank(A^T)?
|
Yes
|
|
Is Ker(A) = Ker (AA^-1)?
|
No, Ker(A) = Ker(AA^T)
|
|
What does it mean if ker(A) = 0?
|
AA^T is invertible
|
|
What are the least squares solutions of Ax=b?
|
x* = (A^TA)^-1 (A^T) (b)
(This is a matrix) |
|
What type of entity is the least squares solution?
|
It is a matrix.
|
|
Describe the determinant in words?
|
The product of the values along the diagnoal.
|
|
What is the form of a discrete trajectory?
|
x(t+1) = Ax(t)
|
|
What is the form of the characteristics equation?
|
det(A-λI)
|
|
What is the characteristic equatin of an n x n matrix?
|
(-λ)^n + tr(A) (-λ)^(n-1) + ... + det(A) = 0
|
|
What is trace?
|
Sum of the diagonal entries.
|
|
What is algebraic multiplicity?
|
# of times a value is a root.
|
|
Do A and A^T have the same eigenvalues?
|
Yes
|
|
Do A and A^ T have the same characteristic polynomials?
|
Yes
|
|
What does it mean that 2 matrices are similar?
|
They have the same det and tr
|
|
How do you find closed formula, (like lake pollutant quesiton)?
|
1. x0 is the initial vector
a b c 2. x0 = c1v1 + c2v2 + ... + cnvn (Find eigenbasis to find c's) 3. x(t) = A^t x0 = c1 λ1^t v1 + ... + cn λn^t vt 4. Plug in values |
|
How do you know if an nxn matrix is diagonizable?
|
If the dimensions of the iegnebasis add up to n
|
|
What does this mean?:
lim A^t t→∞ |
Diagonalize A
|
|
What is the relationsihp between A, S, and D?
|
A = (S^-1) (D) (S)
|
|
How can you find the closed formula solution?
|
x(t) = A^t x0
1. Use D = (S^-1) (A) (S) to find D 2. A^t = (S^-1) (D^t) (S) 3. Since x(t) = (A^t) (x0), x(t) = (S^-1) (D^t) (S) )x0) |
|
When it says find closed formula, whta does it mean in other words?
|
Find all entries of A^t.
|
|
What is DeMoivre's formula?
|
(cosΘ + i sinΘ)^n = cos(nΘ) + i sin(nΘ)
|
|
If a 2 x 2 matrix has eigenvalues of a +ib and a - ib, what is S^-1 A S equal to?
|
The matrix:
a -b b a |
|
If z = a + bi, how do you find r and Θ?
|
r = root (a^2 + b^2)
Θ = arctan (b/a) |
|
What do you do when you want to find some x(t + 1) = Ax(t) fraction as lim goes to infinity?
|
Find eigenbasis, assuming λ = 1, and normalize it.
|
|
When is A orthonogally diagonizable?
|
If it is symmetric, (A^T= A)
|
|
What are 2 ways to model evolution over time?
|
Discrete and continuous cases.
|
|
Describe a discrete case in words.
|
It's like bank accounts tht compound 7% interest each year.
|
|
Describe a discrete case algebraically.
|
X(t+1) = 1.07 x(t)
|
|
Describe a continuous case in words.
|
It's like bank accounts with 7% annual interest, compounded continuously.
|
|
Describe a continuous case algebrically.
|
dx/dt = 0.07 x ~ x = x0 e^(0.07t)
|
|
How do you find solutions to a discrete case, dx/dt = Ax, given A?
|
1. Find the eigenvectors of A and designate this as S.
2. Plug λs and eignebasises into the equation x(t) = c1 e^(λ1t) v1 + ... + cn e^(λnt) vn - You can solve for c's using x0 -The dominant term si associated with the larger eigenvalue, because e is raised to a larger exponent. It, therefore, determines the behavior of the system in the distance future. |
|
Describe the dominant term in a discrete case.
|
the dominant term is associated with the larger eigenvalue, because e is raised to a larger exponent. It, therefore, determine teh behavior of the system in the distant future.
|
|
Rewrite: e^(it)=cos(t) + i sint(t)
|
e^(p+iq)t = e^(pt) (cos(qt) + i sin(qt))
|
|
When is the continuous solution, dx/dt = Ax, a stable equilibrium solution?
|
When all real parts of the eigenvalues are negative.
|
|
What is the zero state of the discrete solution stable?
|
tr(A) is less than 0
det(A) is greater than 0 |
|
What do you do if you are asked to solve a discrete system, where A is a 2x2 matrix with complex eigenvalues?
|
1. Subtract positive λ from the diagonal
2. Find the kernel of this matrix 3. Seperate into a [real coefficients] + i [complex coefficients] - Designate S = [real coefficients complex coefficients] 4. Plug into x(t) = e^(pt) • S • the matrix: cos(qt) -sin(qt) sin(qt) cos(qt) • S^(-1) • x0 |