Chapter 19: covariance matrix
19.1 vector direct product
- scalar = rank-0 tensor
- vector = rank-1 tensor
- matrix = rank-2 tensor
- vector direct product = rank-1 tensor times rank-1 tensor equals rank-2 tensor: increasing rank
- vector inner product = rank-1 tensor times rank-1 tensor equals rank-0 tensor: decreasing rank
scalar = rank-0 tensor
vector = rank-1 tensor
matrix = rank-2 tensor
19.1.1 vector direct product: increasing rank
vector direct product = rank-1 tensor times rank-1 tensor equals rank-2 tensor: increasing rank
U⊗V=UV⊺=UiVj=(U1U2U3)⊗(V1V2V3)=(U1U2U3)(V1V2V3)=(U1V1U1V2U1V3U2V1U2V2U2V3U3V1U3V2U3V3)=(U1V⊺U2V⊺U3V⊺)=(UV1UV2UV3)
19.1.2 vector inner product: decreasing rank
vector inner product = rank-1 tensor times rank-1 tensor equals rank-0 tensor: decreasing rank
U⋅V=V⊺U=ViUi=(U1U2U3)⋅(V1V2V3)=(V1V2V3)(U1U2U3)=V1U1+V2U2+V3U3
19.1.3 tensor direct product: increasing rank
S⊗T=SikTjl(i,j),(k,l)∈{(1,1),(1,2),(2,1),(2,2)}(S11S12S21S22)⊗(T11T12T21T22)=(S11T11S11T12S12T11S12T12S11T21S11T22S12T21S12T22S21T11S21T12S22T11S22T12S21T21S21T22S22T21S22T22)(S11S12S21S22)(T11T12T21T22)=(S11TS12TS21TS22T)=(S11(T11T12T21T22)S12(T11T12T21T22)S21(T11T12T21T22)S22(T11T12T21T22))
19.2 covariance matrix and its properties
C[X]=Cov[X]=V[X]=E[[X−E(X)][X−E(X)]⊺]=E[[X−E(X)][X⊺−E(X)⊺]]=E[XX⊺−E(X)X⊺−XE(X)⊺+E(X)E(X)⊺]=E[XX⊺]−E[E(X)X⊺]−E[XE(X)⊺]+E[E(X)E(X)⊺]=E[XX⊺]−E(X)E[X⊺]−E[X]E(X)⊺+E(X)E(X)⊺=E[XX⊺]−E(X)E(X)⊺−E(X)E(X)⊺+E(X)E(X)⊺=E[XX⊺]−E(X)E(X)⊺
X=[X]1×1=X⇒C(X)=C[X]=E[XX⊺]−E(X)E(X)⊺=E[XX]−E(X)E(X)=E(X2)−[E(X)]2=V(X)
19.2.1 V[X+b]=V[X]
V[X+b]=E[[(X+b)−E(X+b)][(X+b)−E(X+b)]⊺]E(X+b)=E(X)+b=E[[X+b−E(X)−b][X+b−E(X)−b]⊺]=E[[X−E(X)][X−E(X)]⊺]=V[X]