Linear Algebra: Vector Spaces, Eigenvalues, and Orthogonality

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/37

flashcard set

Earn XP

Description and Tags

Comprehensive practice flashcards covering vector spaces, subspaces, null/column/row spaces, dimension and rank, coordinate systems, change of basis, eigenvalues and eigenvectors, similarity, diagonalization, linear transformation matrices, complex eigenvalues, inner products, orthogonality, projections, Gram-Schmidt, QR factorization, and least-squares problems.

38 Terms

1
New cards

What three defining conditions must a subset HH of a vector space VV satisfy to be a subspace?

A subset HH is a subspace if: (1) the zero vector of VV is in HH (0∈H0 \in H); (2) for all u,v∈Hu, v \in H, u+v∈Hu + v \in H (closed under addition); and (3) for each v∈Hv \in H and any scalar cc, cv∈Hcv \in H (closed under scalar multiplication).

2
New cards

If v1,v2,…,vpv_1, v_2, \dots, v_p are vectors in a vector space VV, what theorem provides the main method for producing subspaces of VV?

The theorem states that Span{v1,v2,…,vp}\text{Span}\{v_1, v_2, \dots, v_p\} is always a subspace of VV.

3
New cards

How are the null space, column space, and row space of an m×nm \times n matrix AA defined?

Nul A={x∈Rn:Ax=0}\text{Nul}\,A = \{x \in \mathbb{R}^n : Ax = 0\} is a subspace of Rn\mathbb{R}^n; Col A=Span{a1,…,an}\text{Col}\,A = \text{Span}\{a_1, \dots, a_n\} is a subspace of Rm\mathbb{R}^m; and Row A=Span{r1,…,rm}=Col AT\text{Row}\,A = \text{Span}\{r_1, \dots, r_m\} = \text{Col}\,A^T is a subspace of Rn\mathbb{R}^n.

4
New cards

For a linear transformation T:Rn→RmT: \mathbb{R}^n \to \mathbb{R}^m with standard matrix AA, what are the kernel and range of TT?

The kernel of TT is equal to Nul A\text{Nul}\,A, and the range of TT is equal to Col A\text{Col}\,A.

5
New cards

What two conditions are required for a set of vectors β={b1,…,bp}\beta = \{b_1, \dots, b_p\} to be a basis for a subspace HH?

The set β\beta must be linearly independent, and the span of β\beta must equal HH (Span{b1,…,bp}=H\text{Span}\{b_1, \dots, b_p\} = H).

6
New cards

Given a matrix AA row equivalent to an echelon form BB, how are bases for Row A\text{Row}\,A and Col A\text{Col}\,A determined?

A basis for Row A\text{Row}\,A consists of the nonzero rows of echelon form BB, whereas a basis for Col A\text{Col}\,A consists of the pivot columns of the original matrix AA.

7
New cards

What is the relationship between rank, nullity, and column count given by the Rank-Nullity Theorem for an m×nm \times n matrix AA?

rank A+nullity A=n\text{rank}\,A + \text{nullity}\,A = n, where rank A\text{rank}\,A is the number of pivot positions of AA and nullity A\text{nullity}\,A is the number of free variables in Ax=0Ax = 0.

8
New cards

What does the Basis Theorem state regarding a subset SS of a pp-dimensional vector space VV (p≥1p \ge 1)?

If SS has exactly pp elements, then SS is a basis for VV if either SS is linearly independent or SS spans VV.

9
New cards

What does the Unique Representation Theorem establish about a basis β={b1,…,bn}\beta = \{b_1, \dots, b_n\} for a vector space VV?

For each x∈Vx \in V, there exists a unique set of scalars c1,…,cnc_1, \dots, c_n such that x=c1b1+⋯+cnbnx = c_1b_1 + \dots + c_nb_n.

10
New cards

Under what condition are two finite-dimensional vector spaces VV and WW isomorphic (V≅WV \cong W)?

Two finite-dimensional vector spaces VV and WW are isomorphic if and only if dim⁡V=dim⁡W\dim V = \dim W.

11
New cards

How is the change-of-coordinates matrix PC←BP_{C \leftarrow B} constructed from basis B={b1,…,bn}B = \{b_1, \dots, b_n\} to basis CC?

The columns of PC←BP_{C \leftarrow B} are the CC-coordinate vectors of the vectors in BB: PC←B=[[b1]C  [b2]C  …  [bn]C]P_{C \leftarrow B} = [ [b_1]_C \; [b_2]_C \; \dots \; [b_n]_C ].

12
New cards

In Rn\mathbb{R}^n, what row reduction algorithm computes the change-of-coordinates matrix PC←BP_{C \leftarrow B} from basis BB to basis CC?

Row reduce the double-augmented matrix [c1  …  cn∣b1  …  bn][c_1 \; \dots \; c_n \mid b_1 \; \dots \; b_n] to [In∣PC←B][I_n \mid P_{C \leftarrow B}].

<p>Row reduce the double-augmented matrix $$[c_1 \; \dots \; c_n \mid b_1 \; \dots \; b_n]$$ to $$[I_n \mid P_{C \leftarrow B}]$$.</p>
13
New cards

How can the change-of-coordinates matrix PC←BP_{C \leftarrow B} be calculated using change-of-coordinates matrices PBP_B and PCP_C to the standard basis?

PC←B=PC−1PBP_{C \leftarrow B} = P_C^{-1} P_B, which follows from x=PB[x]Bx = P_B[x]_B and [x]C=PC−1x[x]_C = P_C^{-1}x.

14
New cards

What is the definition of an eigenvector and an eigenvalue of an n×nn \times n matrix AA?

An eigenvector of AA is a nonzero vector x∈Rnx \in \mathbb{R}^n such that Ax=λxAx = \lambda x for some scalar λ\lambda, and λ\lambda is an eigenvalue of AA.

15
New cards

How is the eigenspace of an n×nn \times n matrix AA corresponding to eigenvalue λ\lambda defined?

The eigenspace is Nul(A−λI)\text{Nul}(A - \lambda I), which is the solution set of (A−λI)x=0(A - \lambda I)x = 0, containing the zero vector together with all eigenvectors corresponding to λ\lambda.

16
New cards

What is the relationship between the invertibility of a square matrix AA and the eigenvalue 00?

00 is an eigenvalue of AA if and only if AA is not invertible (which occurs if and only if det⁡A=0\det A = 0).

17
New cards

When are two square matrices AA and BB defined to be similar, and what invariant do they share?

AA and BB are similar if there exists an invertible matrix PP such that P−1AP=BP^{-1}AP = B; similar matrices have identical characteristic polynomials, and hence the same eigenvalues with the same algebraic multiplicities.

18
New cards

According to the Diagonalization Theorem, when is an n×nn \times n matrix AA diagonalizable?

An n×nn \times n matrix AA is diagonalizable if and only if AA has nn linearly independent eigenvectors, meaning A=PDP−1A = PDP^{-1} where the columns of PP are nn linearly independent eigenvectors and the diagonal entries of DD are the corresponding eigenvalues.

19
New cards

If A=PDP−1A = PDP^{-1} is a diagonalizable matrix, how is AkA^k calculated for any positive integer kk?

Ak=PDkP−1A^k = P D^k P^{-1}, where DkD^k is obtained by raising each diagonal entry of DD to the power kk.

20
New cards

What is the formula for the matrix MM of a linear transformation T:V→WT: V \to W relative to bases B={b1,…,bn}B = \{b_1, \dots, b_n\} and CC?

M=[[T(b1)]C  [T(b2)]C  …  [T(bn)]C]M = [ [T(b_1)]_C \; [T(b_2)]_C \; \dots \; [T(b_n)]_C ], satisfying [T(x)]C=M[x]B[T(x)]_C = M [x]_B.

21
New cards

If A=PDP−1A = PDP^{-1} with DD diagonal and BB is the basis formed from the columns of PP, what is DD in terms of linear transformations?

DD is the BB-matrix ([T]B[T]_B) for the transformation x↦Axx \mapsto Ax.

22
New cards

If a real 2×22 \times 2 matrix AA has complex eigenvalue λ=a−bi\lambda = a - bi (b≠0b \neq 0) and complex eigenvector vv, how is AA factored into PCP−1PCP^{-1}?

A=PCP−1A = PCP^{-1}, where P=[Re v    Im v]P = [\text{Re}\,v \;\; \text{Im}\,v] and C=[[a,−b],[b,a]]C = [[a, -b], [b, a]], which represents a rotation and scaling.

23
New cards

Why do the eigenvalues and eigenvectors of a real matrix always occur in complex conjugate pairs?

Because the characteristic polynomial has real coefficients, so roots satisfy p(λˉ)=0p(\bar{\lambda}) = 0 whenever p(λ)=0p(\lambda) = 0, and conjugating Av=λvAv = \lambda v gives Avˉ=λˉvˉA\bar{v} = \bar{\lambda}\bar{v}.

24
New cards

How is the orthogonal complement W⊥W^\perp of a subspace WW of Rn\mathbb{R}^n defined, and what is W∩W⊥W \cap W^\perp?

W⊥={z∈Rn:z⋅w=0 for all w∈W}W^\perp = \{z \in \mathbb{R}^n : z \cdot w = 0 \text{ for all } w \in W\}, and W∩W⊥={0}W \cap W^\perp = \{0\}.

<p>$$W^\perp = \{z \in \mathbb{R}^n : z \cdot w = 0 \text{ for all } w \in W\}$$, and $$W \cap W^\perp = \{0\}$$.</p>
25
New cards

What is the Pythagorean Theorem for vectors uu and vv in Rn\mathbb{R}^n?

uu and vv are orthogonal if and only if ∣∣u+v∣∣2=∣∣u∣∣2+∣∣v∣∣2||u + v||^2 = ||u||^2 + ||v||^2.

26
New cards

What are the orthogonal complements of Row A\text{Row}\,A and Col A\text{Col}\,A for an m×nm \times n matrix AA?

(Row A)⊥=Nul A(\text{Row}\,A)^\perp = \text{Nul}\,A in Rn\mathbb{R}^n, and (Col A)⊥=Nul(AT)(\text{Col}\,A)^\perp = \text{Nul}(A^T) in Rm\mathbb{R}^m.

27
New cards

Given an orthogonal basis β={u1,…,up}\beta = \{u_1, \dots, u_p\} for WW, how is each coordinate cjc_j of a vector y∈Wy \in W computed?

cj=y⋅ujuj⋅uj=y⋅uj∣∣uj∣∣2c_j = \frac{y \cdot u_j}{u_j \cdot u_j} = \frac{y \cdot u_j}{||u_j||^2} for each j=1,…,pj = 1, \dots, p.

28
New cards

What geometric quantities are preserved by a linear transformation defined by an m×nm \times n matrix UU with orthonormal columns?

Multiplication by UU preserves lengths (∣∣Ux∣∣=∣∣x∣∣||Ux|| = ||x||), dot products ((Ux)⋅(Uy)=x⋅y(Ux) \cdot (Uy) = x \cdot y), and orthogonality ((Ux)⋅(Uy)=0  ⟺  x⋅y=0(Ux) \cdot (Uy) = 0 \iff x \cdot y = 0).

29
New cards

What does the Orthogonal Decomposition Theorem state for a subspace WW of Rn\mathbb{R}^n and any vector y∈Rny \in \mathbb{R}^n?

Each y∈Rny \in \mathbb{R}^n can be written uniquely as y=y^+zy = \hat{y} + z, where y^=projWy∈W\hat{y} = \text{proj}_W y \in W and z=y−y^∈W⊥z = y - \hat{y} \in W^\perp.

30
New cards

If the columns of an n×pn \times p matrix UU form an orthonormal basis for a subspace WW, what matrix represents the orthogonal projection onto WW?

The standard matrix is UUTUU^T, meaning projWy=UUTy\text{proj}_W y = UU^T y.

31
New cards

What does the Best Approximation Theorem state regarding the orthogonal projection projWy\text{proj}_W y?

projWy\text{proj}_W y is the closest point in WW to yy, such that ∣∣y−projWy∣∣<∣∣y−v∣∣||y - \text{proj}_W y|| < ||y - v|| for any v∈Wv \in W with v≠projWyv \neq \text{proj}_W y.

32
New cards

What is the general formula for each vector vkv_k in the Gram–Schmidt process given a basis \Base = \{x_1, \dots, x_p\}?

v1=x1v_1 = x_1, and for k≥2k \ge 2, vk=xk−xk⋅v1v1⋅v1v1−⋯−xk⋅vk−1vk−1⋅vk−1vk−1v_k = x_k - \frac{x_k \cdot v_1}{v_1 \cdot v_1}v_1 - \dots - \frac{x_k \cdot v_{k-1}}{v_{k-1} \cdot v_{k-1}}v_{k-1}.

33
New cards

In the QR decomposition A=QRA = QR for an m×nm \times n matrix AA with linearly independent columns, what are the properties of QQ and RR, and how is RR computed?

QQ is an m×nm \times n matrix with orthonormal columns, RR is an n×nn \times n upper triangular invertible matrix with positive diagonal entries, and R=QTAR = Q^T A.

34
New cards

What is the definition of a least-squares solution to an inconsistent system Ax=bAx = b?

A least-squares solution is a vector x^∈Rn\hat{x} \in \mathbb{R}^n such that ∣∣b−Ax^∣∣≤∣∣b−Ax∣∣||b - A\hat{x}|| \le ||b - Ax|| for all x∈Rnx \in \mathbb{R}^n, satisfying Ax^=projCol AbA\hat{x} = \text{proj}_{\text{Col}\,A} b.

35
New cards

What matrix equation defines the normal equations for finding a least-squares solution to Ax=bAx = b, and what is the formula when ATAA^T A is invertible?

The normal equations are ATAx=ATbA^T Ax = A^T b, and when the columns of AA are linearly independent, the unique least-squares solution is x^=(ATA)−1ATb\hat{x} = (A^T A)^{-1}A^T b.

36
New cards

How can the QR factorization A=QRA = QR be used to find the least-squares solution to Ax=bAx = b?

The least-squares solution is given by x^=R−1QTb\hat{x} = R^{-1}Q^T b, which can be computed by solving the upper triangular system Rx^=QTbR\hat{x} = Q^T b.

37
New cards

Why can an arbitrary matrix AA NOT be row reduced to echelon form to find its eigenvalues?

Row reduction operations do not preserve eigenvalues; the resulting upper triangular echelon matrix is generally not similar to AA and has different eigenvalues.

38
New cards

How does the coordinate mapping x↦[x]Bx \mapsto [x]_B preserve algebraic structure between an nn-dimensional vector space VV and Rn\mathbb{R}^n?

The coordinate mapping is a one-to-one and onto linear transformation (an isomorphism) that maps basis vectors \Base = \{b_1, \dots, b_n\} to the standard basis vectors \Base_E = \{e_1, \dots, e_n\}.