The Matrix Cookbook

[ http://matrixcookbook.com ]

Kaare Brandt Petersen

Michael Syskind Pedersen
Version: November 15, 2012

1
Introduction
What is this? These pages are a collection of facts (identities, approxima-
tions, inequalities, relations, ...) about matrices and matters relating to them.
It is collected in this form for the convenience of anyone who wants a quick
desktop reference .

Disclaimer: The identities, approximations and relations presented here were

obviously not invented but collected, borrowed and copied from a large amount
of sources. These sources include similar but shorter notes found on the internet
and appendices in books - see the references for a full list.

Errors: Very likely there are errors, typos, and mistakes for which we apolo-
gize and would be grateful to receive corrections at [email protected].

Its ongoing: The project of keeping a large repository of relations involving

matrices is naturally ongoing and the version will be apparent from the date in
the header.

Suggestions: Your suggestion for additional content or elaboration of some

topics is most welcome [email protected].

Keywords: Matrix algebra, matrix relations, matrix identities, derivative of

determinant, derivative of inverse matrix, differentiate a matrix.

Acknowledgements: We would like to thank the following for contributions

and suggestions: Bill Baxter, Brian Templeton, Christian Rishj, Christian
Schroppel, Dan Boley, Douglas L. Theobald, Esben Hoegh-Rasmussen, Evripidis
Karseras, Georg Martius, Glynne Casteel, Jan Larsen, Jun Bin Gao, Jurgen
Struckmeier, Kamil Dedecius, Karim T. Abou-Moustafa, Korbinian Strimmer,
Lars Christiansen, Lars Kai Hansen, Leland Wilkinson, Liguo He, Loic Thibaut,
Markus Froeb, Michael Hubatka, Miguel Barao, Ole Winther, Pavel Sakov,
Stephan Hattinger, Troels Pedersen, Vasile Sima, Vincent Rabaud, Zhaoshui
He. We would also like thank The Oticon Foundation for funding our PhD
studies.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 2
CONTENTS CONTENTS

Contents
1 Basics 6
1.1 Trace . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.2 Determinant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
1.3 The Special Case 2x2 . . . . . . . . . . . . . . . . . . . . . . . . . 7

2 Derivatives 8
2.1 Derivatives of a Determinant . . . . . . . . . . . . . . . . . . . . 8
2.2 Derivatives of an Inverse . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Derivatives of Eigenvalues . . . . . . . . . . . . . . . . . . . . . . 10
2.4 Derivatives of Matrices, Vectors and Scalar Forms . . . . . . . . 10
2.5 Derivatives of Traces . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.6 Derivatives of vector norms . . . . . . . . . . . . . . . . . . . . . 14
2.7 Derivatives of matrix norms . . . . . . . . . . . . . . . . . . . . . 14
2.8 Derivatives of Structured Matrices . . . . . . . . . . . . . . . . . 14

3 Inverses 17
3.1 Basic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.2 Exact Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.3 Implication on Inverses . . . . . . . . . . . . . . . . . . . . . . . . 20
3.4 Approximations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
3.5 Generalized Inverse . . . . . . . . . . . . . . . . . . . . . . . . . . 21
3.6 Pseudo Inverse . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

4 Complex Matrices 24
4.1 Complex Derivatives . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2 Higher order and non-linear derivatives . . . . . . . . . . . . . . . 26
4.3 Inverse of complex sum . . . . . . . . . . . . . . . . . . . . . . . 27

5 Solutions and Decompositions 28

5.1 Solutions to linear equations . . . . . . . . . . . . . . . . . . . . . 28
5.2 Eigenvalues and Eigenvectors . . . . . . . . . . . . . . . . . . . . 30
5.3 Singular Value Decomposition . . . . . . . . . . . . . . . . . . . . 31
5.4 Triangular Decomposition . . . . . . . . . . . . . . . . . . . . . . 32
5.5 LU decomposition . . . . . . . . . . . . . . . . . . . . . . . . . . 32
5.6 LDM decomposition . . . . . . . . . . . . . . . . . . . . . . . . . 33
5.7 LDL decompositions . . . . . . . . . . . . . . . . . . . . . . . . . 33

6 Statistics and Probability 34

6.1 Definition of Moments . . . . . . . . . . . . . . . . . . . . . . . . 34
6.2 Expectation of Linear Combinations . . . . . . . . . . . . . . . . 35
6.3 Weighted Scalar Variable . . . . . . . . . . . . . . . . . . . . . . 36

7 Multivariate Distributions 37
7.1 Cauchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.2 Dirichlet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.3 Normal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.4 Normal-Inverse Gamma . . . . . . . . . . . . . . . . . . . . . . . 37
7.5 Gaussian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.6 Multinomial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 3
CONTENTS CONTENTS

7.7 Students t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
7.8 Wishart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
7.9 Wishart, Inverse . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

8 Gaussians 40
8.1 Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
8.2 Moments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
8.3 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
8.4 Mixture of Gaussians . . . . . . . . . . . . . . . . . . . . . . . . . 44

9 Special Matrices 46
9.1 Block matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
9.2 Discrete Fourier Transform Matrix, The . . . . . . . . . . . . . . 47
9.3 Hermitian Matrices and skew-Hermitian . . . . . . . . . . . . . . 48
9.4 Idempotent Matrices . . . . . . . . . . . . . . . . . . . . . . . . . 49
9.5 Orthogonal matrices . . . . . . . . . . . . . . . . . . . . . . . . . 49
9.6 Positive Definite and Semi-definite Matrices . . . . . . . . . . . . 50
9.7 Singleentry Matrix, The . . . . . . . . . . . . . . . . . . . . . . . 52
9.8 Symmetric, Skew-symmetric/Antisymmetric . . . . . . . . . . . . 54
9.9 Toeplitz Matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
9.10 Transition matrices . . . . . . . . . . . . . . . . . . . . . . . . . . 55
9.11 Units, Permutation and Shift . . . . . . . . . . . . . . . . . . . . 56
9.12 Vandermonde Matrices . . . . . . . . . . . . . . . . . . . . . . . . 57

10 Functions and Operators 58

10.1 Functions and Series . . . . . . . . . . . . . . . . . . . . . . . . . 58
10.2 Kronecker and Vec Operator . . . . . . . . . . . . . . . . . . . . 59
10.3 Vector Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.4 Matrix Norms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
10.5 Rank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
10.6 Integral Involving Dirac Delta Functions . . . . . . . . . . . . . . 62
10.7 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

A One-dimensional Results 64
A.1 Gaussian . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
A.2 One Dimensional Mixture of Gaussians . . . . . . . . . . . . . . . 65

B Proofs and Details 66

B.1 Misc Proofs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 4
CONTENTS CONTENTS

Notation and Nomenclature

A Matrix
Aij Matrix indexed for some purpose
Ai Matrix indexed for some purpose
Aij Matrix indexed for some purpose
An Matrix indexed for some purpose or
The n.th power of a square matrix
A1 The inverse matrix of the matrix A
A+ The pseudo inverse matrix of the matrix A (see Sec. 3.6)
A1/2 The square root of a matrix (if unique), not elementwise
(A)ij The (i, j).th entry of the matrix A
Aij The (i, j).th entry of the matrix A
[A]ij The ij-submatrix, i.e. A with i.th row and j.th column deleted
a Vector (column-vector)
ai Vector indexed for some purpose
ai The i.th element of the vector a
a Scalar

<z Real part of a scalar

<z Real part of a vector
<Z Real part of a matrix
=z Imaginary part of a scalar
=z Imaginary part of a vector
=Z Imaginary part of a matrix

det(A) Determinant of A
Tr(A) Trace of the matrix A
diag(A) Diagonal matrix of the matrix A, i.e. (diag(A))ij = ij Aij
eig(A) Eigenvalues of the matrix A
vec(A) The vector-version of the matrix A (see Sec. 10.2.2)
sup Supremum of a set
||A|| Matrix norm (subscript if any denotes what norm)
AT Transposed matrix
AT The inverse of the transposed and vice versa, AT = (A1 )T = (AT )1 .
A Complex conjugated matrix
AH Transposed and complex conjugated matrix (Hermitian)

AB Hadamard (elementwise) product

AB Kronecker product

0 The null matrix. Zero in all entries.

I The identity matrix
Jij The single-entry matrix, 1 at (i, j) and zero elsewhere
A positive definite matrix
A diagonal matrix

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 5
 1 BASICS

1 Basics

(AB)1 = B1 A1 (1)
1 1 1 1
(ABC...) = ...C B A (2)
T 1 1 T
(A ) = (A ) (3)
T T T
(A + B) = A +B (4)
(AB)T = BT AT (5)
(ABC...)T = ...CT BT AT (6)
H 1 1 H
(A ) = (A ) (7)
(A + B)H = AH + BH (8)
(AB)H = BH AH (9)
H H H H
(ABC...) = ...C B A (10)

1.1 Trace
P
Tr(A) = Aii (11)
Pi
Tr(A) = i i , i = eig(A) (12)
T
Tr(A) = Tr(A ) (13)
Tr(AB) = Tr(BA) (14)
Tr(A + B) = Tr(A) + Tr(B) (15)
Tr(ABC) = Tr(BCA) = Tr(CAB) (16)
aT a = Tr(aaT ) (17)

1.2 Determinant
Let A be an n n matrix.
Q
det(A) = i i i = eig(A) (18)
det(cA) = cn det(A), if A Rnn (19)
det(AT ) = det(A) (20)
det(AB) = det(A) det(B) (21)
det(A1 ) = 1/ det(A) (22)
det(An ) = det(A)n (23)
T T
det(I + uv ) = 1+u v (24)

For n = 2:
det(I + A) = 1 + det(A) + Tr(A) (25)
For n = 3:
1 1
det(I + A) = 1 + det(A) + Tr(A) + Tr(A)2 Tr(A2 ) (26)
2 2

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 6
1.3 The Special Case 2x2 1 BASICS

For n = 4:
1
det(I + A) = 1 + det(A) + Tr(A) +
2
1
+Tr(A)2 Tr(A2 )
2
1 1 1
+ Tr(A) Tr(A)Tr(A2 ) + Tr(A3 )
3
(27)
6 2 3
For small , the following approximation holds
1 1
det(I + A)
= 1 + det(A) + Tr(A) + 2 Tr(A)2 2 Tr(A2 ) (28)
2 2

1.3 The Special Case 2x2

Consider the matrix A  
A11 A12
A=
A21 A22
Determinant and trace

det(A) = A11 A22 A12 A21 (29)

Tr(A) = A11 + A22 (30)

Eigenvalues
2 Tr(A) + det(A) = 0
p p
Tr(A) + Tr(A)2 4 det(A) Tr(A) Tr(A)2 4 det(A)
1 = 2 =
2 2
1 + 2 = Tr(A) 1 2 = det(A)
Eigenvectors    
A12 A12
v1 v2
1 A11 2 A11
Inverse  
1 1 A22 A12
A = (31)
det(A) A21 A11

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 7
 2 DERIVATIVES

2 Derivatives
This section is covering differentiation of a number of expressions with respect to
a matrix X. Note that it is always assumed that X has no special structure, i.e.
that the elements of X are independent (e.g. not symmetric, Toeplitz, positive
definite). See section 2.8 for differentiation of structured matrices. The basic
assumptions can be written in a formula as
Xkl
= ik lj (32)
Xij

that is for e.g. vector forms,

     
x xi x x x xi
= = =
y i y y i yi y ij yj

The following rules are general and very useful when deriving the differential of
an expression ([19]):

A = 0 (A is a constant) (33)
(X) = X (34)
(X + Y) = X + Y (35)
(Tr(X)) = Tr(X) (36)
(XY) = (X)Y + X(Y) (37)
(X Y) = (X) Y + X (Y) (38)
(X Y) = (X) Y + X (Y) (39)
(X1 ) = X1 (X)X1 (40)
(det(X)) = Tr(adj(X)X) (41)
(det(X)) = det(X)Tr(X1 X) (42)
(ln(det(X))) = Tr(X1 X) (43)
XT = (X)T (44)
XH = (X)H (45)

2.1 Derivatives of a Determinant

2.1.1 General form
 
det(Y) 1 Y
= det(Y)Tr Y (46)
x x
X det(X)
Xjk = ij det(X) (47)
Xik
k
" " #
Y
2 det(Y) 1 x
= det(Y) Tr Y
x2 x
   
Y Y
+Tr Y1 Tr Y1
x x
   #
1 Y 1 Y
Tr Y Y (48)
x x

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 8
2.2 Derivatives of an Inverse 2 DERIVATIVES

2.1.2 Linear forms

det(X)
= det(X)(X1 )T (49)
X
X det(X)
Xjk = ij det(X) (50)
Xik
k
det(AXB)
= det(AXB)(X1 )T = det(AXB)(XT )1 (51)
X

2.1.3 Square forms

If X is square and invertible, then

det(XT AX)
= 2 det(XT AX)XT (52)
X
If X is not square but A is symmetric, then

det(XT AX)
= 2 det(XT AX)AX(XT AX)1 (53)
X
If X is not square and A is not symmetric, then

det(XT AX)
= det(XT AX)(AX(XT AX)1 + AT X(XT AT X)1 ) (54)
X

2.1.4 Other nonlinear forms

Some special cases are (See [9, 7])

ln det(XT X)|
= 2(X+ )T (55)
X
ln det(XT X)
= 2XT (56)
X+
ln | det(X)|
= (X1 )T = (XT )1 (57)
X
det(Xk )
= k det(Xk )XT (58)
X

2.2 Derivatives of an Inverse

From [27] we have the basic identity

Y1 Y 1
= Y1 Y (59)
x x

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 9
2.3 Derivatives of Eigenvalues 2 DERIVATIVES

from which it follows

(X1 )kl
= (X1 )ki (X1 )jl (60)
Xij
aT X1 b
= XT abT XT (61)
X
det(X1 )
= det(X1 )(X1 )T (62)
X
Tr(AX1 B)
= (X1 BAX1 )T (63)
X
Tr((X + A)1 )
= ((X + A)1 (X + A)1 )T (64)
X
From [32] we have the following result: Let A be an n n invertible square
matrix, W be the inverse of A, and J(A) is an n n -variate and differentiable
function with respect to A, then the partial differentials of J with respect to A
and W satisfy
J J T
= AT A
A W

2.3 Derivatives of Eigenvalues

X
eig(X) = Tr(X) = I (65)
X X
Y
eig(X) = det(X) = det(X)XT (66)
X X
If A is real and symmetric, i and vi are distinct eigenvalues and eigenvectors
of A (see (276)) with viT vi = 1, then [33]
i = viT (A)vi (67)
+
vi = (i I A) (A)vi (68)

2.4 Derivatives of Matrices, Vectors and Scalar Forms

2.4.1 First Order

xT a aT x
= = a (69)
x x
aT Xb
= abT (70)
X
aT XT b
= baT (71)
X
aT Xa aT XT a
= = aaT (72)
X X
X
= Jij (73)
Xij
(XA)ij
= im (A)nj = (Jmn A)ij (74)
Xmn
(XT A)ij
= in (A)mj = (Jnm A)ij (75)
Xmn

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 10
2.4 Derivatives of Matrices, Vectors and Scalar Forms 2 DERIVATIVES

2.4.2 Second Order

X X
Xkl Xmn = 2 Xkl (76)
Xij
klmn kl
bT XT Xc
= X(bcT + cbT ) (77)
X
(Bx + b)T C(Dx + d)
= BT C(Dx + d) + DT CT (Bx + b) (78)
x
(XT BX)kl
= lj (XT B)ki + kj (BX)il (79)
Xij
(XT BX)
= XT BJij + Jji BX (Jij )kl = ik jl (80)
Xij

See Sec 9.7 for useful properties of the Single-entry matrix Jij

xT Bx
= (B + BT )x (81)
x
bT XT DXc
= DT XbcT + DXcbT (82)
X

(Xb + c)T D(Xb + c) = (D + DT )(Xb + c)bT (83)
X
Assume W is symmetric, then

(x As)T W(x As) = 2AT W(x As) (84)
s

(x s)T W(x s) = 2W(x s) (85)
x

(x s)T W(x s) = 2W(x s) (86)
s

(x As)T W(x As) = 2W(x As) (87)
x

(x As)T W(x As) = 2W(x As)sT (88)
A
As a case with complex values the following holds

(a xH b)2
= 2b(a xH b) (89)
x
This formula is also known from the LMS algorithm [14]

2.4.3 Higher-order and non-linear

n1
(Xn )kl X
= (Xr Jij Xn1r )kl (90)
Xij r=0

For proof of the above, see B.1.3.

n1
T n X
a X b= (Xr )T abT (Xn1r )T (91)
X r=0

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 11
2.5 Derivatives of Traces 2 DERIVATIVES

n1
T n T n Xh
a (X ) X b = Xn1r abT (Xn )T Xr
X r=0
i
+(Xr )T Xn abT (Xn1r )T (92)

See B.1.3 for a proof.

Assume s and r are functions of x, i.e. s = s(x), r = r(x), and that A is a
constant, then
 T  T
T s r
s Ar = Ar + AT s (93)
x x x

(Ax)T (Ax) xT AT Ax
= (94)
x (Bx)T (Bx) x xT BT Bx
AT Ax xT AT AxBT Bx
= 2 T 2 (95)
x BBx (xT BT Bx)2

2.4.4 Gradient and Hessian

Using the above we have for the gradient and the Hessian
f = xT Ax + bT x (96)
f
x f = = (A + AT )x + b (97)
x
2f
= A + AT (98)
xxT

2.5 Derivatives of Traces

Assume F (X) to be a differentiable function of each of the elements of X. It
then holds that
Tr(F (X))
= f (X)T
X
where f () is the scalar derivative of F ().

2.5.1 First Order

Tr(X) = I (99)
X

Tr(XA) = AT (100)
X

Tr(AXB) = AT BT (101)
X

Tr(AXT B) = BA (102)
X

Tr(XT A) = A (103)
X

Tr(AXT ) = A (104)
X

Tr(A X) = Tr(A)I (105)
X

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 12
2.5 Derivatives of Traces 2 DERIVATIVES

2.5.2 Second Order

Tr(X2 ) = 2XT (106)
X

Tr(X2 B) = (XB + BX)T (107)
X

Tr(XT BX) = BX + BT X (108)
X

Tr(BXXT ) = BX + BT X (109)
X

Tr(XXT B) = BX + BT X (110)
X

Tr(XBXT ) = XBT + XB (111)
X

Tr(BXT X) = XBT + XB (112)
X

Tr(XT XB) = XBT + XB (113)
X

Tr(AXBX) = AT XT BT + BT XT AT (114)
X

Tr(XT X) = Tr(XXT ) = 2X (115)
X X

Tr(BT XT CXB) = CT XBBT + CXBBT (116)
X

Tr XT BXC = BXC + BT XCT
 
(117)
X

Tr(AXBXT C) = AT CT XBT + CAXB (118)
X
h i
Tr (AXB + C)(AXB + C)T = 2AT (AXB + C)BT (119)
X

Tr(X X) = Tr(X)Tr(X) = 2Tr(X)I(120)
X X
See [7].

2.5.3 Higher Order

Tr(Xk ) = k(Xk1 )T (121)
X
k1
X
Tr(AXk ) = (Xr AXkr1 )T (122)
X r=0
 T T T

= CXXT CXBBT

X Tr B X CXX CXB
+CT XBBT XT CT X
+CXBBT XT CX
+CT XXT CT XBBT (123)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 13
2.6 Derivatives of vector norms 2 DERIVATIVES

2.5.4 Other

Tr(AX1 B) = (X1 BAX1 )T = XT AT BT XT (124)
X
Assume B and C to be symmetric, then
h i
Tr (XT CX)1 A = (CX(XT CX)1 )(A + AT )(XT CX)1 (125)
X
h i
Tr (XT CX)1 (XT BX) = 2CX(XT CX)1 XT BX(XT CX)1
X
+2BX(XT CX)1 (126)
h i
Tr (A + XT CX)1 (XT BX) = 2CX(A + XT CX)1 XT BX(A + XT CX)1
X
+2BX(A + XT CX)1 (127)

See [7].

Tr(sin(X))
= cos(X)T (128)
X

2.6 Derivatives of vector norms

2.6.1 Two-norm
xa
||x a||2 = (129)
x ||x a||2
xa I (x a)(x a)T
= (130)
x kx ak2 kx ak2 kx ak32
||x||22 ||xT x||2
= = 2x (131)
x x

2.7 Derivatives of matrix norms

For more on matrix norms, see Sec. 10.4.

2.7.1 Frobenius norm

||X||2F = Tr(XXH ) = 2X (132)
X X
See (248). Note that this is also a special case of the result in equation 119.

2.8 Derivatives of Structured Matrices

Assume that the matrix A has some structure, i.e. symmetric, toeplitz, etc.
In that case the derivatives of the previous section does not apply in general.
Instead, consider the following general rule for differentiating a scalar function
f (A) " #
T
df X f Akl f A
= = Tr (133)
dAij Akl Aij A Aij
kl

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 14
2.8 Derivatives of Structured Matrices 2 DERIVATIVES

The matrix differentiated with respect to itself is in this document referred to

as the structure matrix of A and is defined simply by
A
= Sij (134)
Aij

If A has no special structure we have simply Sij = Jij , that is, the structure
matrix is simply the single-entry matrix. Many structures have a representation
in singleentry matrices, see Sec. 9.7.6 for more examples of structure matrices.

2.8.1 The Chain Rule

Sometimes the objective is to find the derivative of a matrix which is a function
of another matrix. Let U = f (X), the goal is to find the derivative of the
function g(U) with respect to X:

g(U) g(f (X))

= (135)
X X
Then the Chain Rule can then be written the following way:
M N
g(U) g(U) X X g(U) ukl
= = (136)
X xij ukl xij
k=1 l=1

Using matrix notation, this can be written as:

g(U) h g(U) U i
= Tr ( )T . (137)
Xij U Xij

2.8.2 Symmetric
If A is symmetric, then Sij = Jij + Jji Jij Jij and therefore
   T  
df f f f
= + diag (138)
dA A A A

That is, e.g., ([5]):

Tr(AX)
= A + AT (A I), see (142) (139)
X
det(X)
= det(X)(2X1 (X1 I)) (140)
X
ln det(X)
= 2X1 (X1 I) (141)
X

2.8.3 Diagonal
If X is diagonal, then ([19]):

Tr(AX)
= AI (142)
X

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 15
2.8 Derivatives of Structured Matrices 2 DERIVATIVES

2.8.4 Toeplitz
Like symmetric matrices and diagonal matrices also Toeplitz matrices has a
special structure which should be taken into account when the derivative with
respect to a matrix with Toeplitz structure.

Tr(AT)
(143)
T
Tr(TA)
=
T
Tr([AT ]n1 ) Tr([[AT ]1n ]n1,2 )

Tr(A) An1

. . .
. . .
Tr([AT ]1n ))
Tr(A) . . .


. . .
=

. . .
Tr([[AT ]1n ]2,n1 ) . . . Tr([[AT ]1n ]n1,2 )


. . . .
. . . .
. . . . Tr([AT ]n1 )
A1n Tr([[AT ]1n ]2,n1 ) Tr([AT ]1n )) Tr(A)

(A)

As it can be seen, the derivative (A) also has a Toeplitz structure. Each value
in the diagonal is the sum of all the diagonal valued in A, the values in the
diagonals next to the main diagonal equal the sum of the diagonal next to the
main diagonal in AT . This result is only valid for the unconstrained Toeplitz
matrix. If the Toeplitz matrix also is symmetric, the same derivative yields

Tr(AT) Tr(TA)
= = (A) + (A)T (A) I (144)
T T

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 16
 3 INVERSES

3 Inverses
3.1 Basic
3.1.1 Definition
The inverse A1 of a matrix A Cnn is defined such that

AA1 = A1 A = I, (145)

where I is the n n identity matrix. If A1 exists, A is said to be nonsingular.

Otherwise, A is said to be singular (see e.g. [12]).

3.1.2 Cofactors and Adjoint

The submatrix of a matrix A, denoted by [A]ij is a (n 1) (n 1) matrix
obtained by deleting the ith row and the jth column of A. The (i, j) cofactor
of a matrix is defined as

cof(A, i, j) = (1)i+j det([A]ij ), (146)

The matrix of cofactors can be created from the cofactors

cof(A, 1, 1) cof(A, 1, n)


cof(A) =
.. ..
(147)
. cof(A, i, j) .


cof(A, n, 1) cof(A, n, n)

The adjoint matrix is the transpose of the cofactor matrix

adj(A) = (cof(A))T , (148)

3.1.3 Determinant
The determinant of a matrix A Cnn is defined as (see [12])
n
X
det(A) = (1)j+1 A1j det ([A]1j ) (149)
j=1
Xn
= A1j cof(A, 1, j). (150)
j=1

3.1.4 Construction
The inverse matrix can be constructed, using the adjoint matrix, by
1
A1 = adj(A) (151)
det(A)

For the case of 2 2 matrices, see section 1.3.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 17
3.2 Exact Relations 3 INVERSES

3.1.5 Condition number

The condition number of a matrix c(A) is the ratio between the largest and the
smallest singular value of a matrix (see Section 5.3 on singular values),

d+
c(A) = (152)
d

The condition number can be used to measure how singular a matrix is. If the
condition number is large, it indicates that the matrix is nearly singular. The
condition number can also be estimated from the matrix norms. Here

c(A) = kAk kA1 k, (153)

where k k is a norm such as e.g the 1-norm, the 2-norm, the -norm or the
Frobenius norm (see Sec 10.4p for more on matrix norms).
The 2-norm of A equals (max(eig(AH A))) [12, p.57]. For a symmetric
matrix, this reduces to ||A||2 = max(|eig(A)|) [12, p.394]. If the matrix is
symmetric and positive definite, ||A||2 = max(eig(A)). The condition number
based on the 2-norm thus reduces to
max(eig(A))
kAk2 kA1 k2 = max(eig(A)) max(eig(A1 )) = . (154)
min(eig(A))

3.2 Exact Relations

3.2.1 Basic
(AB)1 = B1 A1 (155)

3.2.2 The Woodbury identity

The Woodbury identity comes in many variants. The latter of the two can be
found in [12]

(A + CBCT )1 = A1 A1 C(B1 + CT A1 C)1 CT A1 (156)

1
(A + UBV) = A1 A1 U(B1 + VA1 U)1 VA1 (157)

If P, R are positive definite, then (see [30])

(P1 + BT R1 B)1 BT R1 = PBT (BPBT + R)1 (158)

3.2.3 The Kailath Variant

(A + BC)1 = A1 A1 B(I + CA1 B)1 CA1 (159)
See [4, page 153].

3.2.4 Sherman-Morrison
A1 bcT A1
(A + bcT )1 = A1 (160)
1 + cT A1 b

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 18
3.2 Exact Relations 3 INVERSES

3.2.5 The Searle Set of Identities

The following set of identities, can be found in [25, page 151],
(I + A1 )1 = A(A + I)1 (161)
T 1 1 T 1 1
(A + BB ) B = A B(I + B A B) (162)
(A1 + B1 )1 = A(A + B)1 B = B(A + B)1 A (163)
1 1
A A(A + B) A = B B(A + B) B (164)
1 1 1 1
A +B = A (A + B)B (165)
1 1
(I + AB) = I A(I + BA) B (166)
(I + AB)1 A = A(I + BA)1 (167)

3.2.6 Rank-1 update of inverse of inner product

Denote A = (XT X)1 and that X is extended to include a new column vector
in the end X = [X v]. Then [34]
" T
#
vvT XAT AXT v
T 1 A + vAXT vvT XAXT v T T
v vv XAX v T
(X X) = vT XAT 1
vT vvT XAXT v vT vvT XAXT v

3.2.7 Rank-1 update of Moore-Penrose Inverse

The following is a rank-1 update for the Moore-Penrose pseudo-inverse of real
valued matrices and proof can be found in [18]. The matrix G is defined below:
(A + cdT )+ = A+ + G (168)
Using the the notation
= 1 + dT A+ c (169)
+
v = A c (170)
n = (A+ )T d (171)
w = (I AA+ )c (172)
+ T
m = (I A A) d (173)
the solution is given as six different cases, depending on the entities ||w||,
||m||, and . Please note, that for any (column) vector v it holds that v+ =
vT
vT (vT v)1 = ||v||2 . The solution is:

Case 1 of 6: If ||w|| =
6 0 and ||m|| =
6 0. Then
G = vw+ (m+ )T nT + (m+ )T w+ (174)
1 1
= vwT mnT + mwT (175)
||w||2 ||m||2 ||m||2 ||w||2
Case 2 of 6: If ||w|| = 0 and ||m|| =
6 0 and = 0. Then
G = vv+ A+ (m+ )T nT (176)
1 1
= 2
vvT A+ mnT (177)
||v|| ||m||2

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 19
3.3 Implication on Inverses 3 INVERSES

Case 3 of 6: If ||w|| = 0 and 6= 0. Then

T
||v||2 ||m||2 + T
 
1
G = mvT A+ m+v (A ) v + n
||v||2 ||m||2 + ||2
(178)
Case 4 of 6: If ||w|| =
6 0 and ||m|| = 0 and = 0. Then

G = A+ nn+ vw+ (179)

1 1
= A+ nnT vwT (180)
||n||2 ||w||2
Case 5 of 6: If ||m|| = 0 and 6= 0. Then
T
||w||2 + ||n||2
 
1 +
G= A nwT A n+v w+n (181)
||n||2 ||w||2 + ||2
Case 6 of 6: If ||w|| = 0 and ||m|| = 0 and = 0. Then

G = vv+ A+ A+ nn+ + v+ A+ nvn+ (182)

1 1 v T A+ n
= 2
vvT A+ 2
A+ nnT + vnT (183)
||v|| ||n|| ||v||2 ||n||2

3.3 Implication on Inverses

If (A + B)1 = A1 + B1 then AB1 A = BA1 B (184)
See [25].

3.3.1 A PosDef identity

Assume P, R to be positive definite and invertible, then

(P1 + BT R1 B)1 BT R1 = PBT (BPBT + R)1 (185)

See [30].

3.4 Approximations
The following identity is known as the Neuman series of a matrix, which holds
when |i | < 1 for all eigenvalues i

X
(I A)1 = An (186)
n=0

which is equivalent to

X
(I + A)1 = (1)n An (187)
n=0

When |i | < 1 for all eigenvalues i , it holds that A 0 for n , and the
following approximations holds

(I A)1
= I + A + A2 (188)
(I + A)1
= I A + A2 (189)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 20
3.5 Generalized Inverse 3 INVERSES

The following approximation is from [22] and holds when A large and symmetric

A A(I + A)1 A
= I A1 (190)

If 2 is small compared to Q and M then

(Q + 2 M)1
= Q1 2 Q1 MQ1 (191)

Proof:

(Q + 2 M)1 = (192)
1 2 1 1
(QQ Q + MQ Q) = (193)
((I + 2 MQ1 )Q)1 = (194)
1 2 1 1
Q (I + MQ ) (195)

This can be rewritten using the Taylor expansion:

Q1 (I + 2 MQ1 )1 = (196)
1
Q 2
(I MQ1 2
+ ( MQ 1 2
) ...)
= 1
Q 2
Q 1 1
MQ (197)

3.5 Generalized Inverse

3.5.1 Definition
A generalized inverse matrix of the matrix A is any matrix A such that (see
[26])
AA A = A (198)
The matrix A is not unique.

3.6 Pseudo Inverse

3.6.1 Definition
The pseudo inverse (or Moore-Penrose inverse) of a matrix A is the matrix A+
that fulfils

I AA+ A = A
II A+ AA+ = A+
III AA+ symmetric
IV A+ A symmetric

The matrix A+ is unique and does always exist. Note that in case of com-
plex matrices, the symmetric condition is substituted by a condition of being
Hermitian.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 21
3.6 Pseudo Inverse 3 INVERSES

3.6.2 Properties
Assume A+ to be the pseudo-inverse of A, then (See [3] for some of them)

(A+ )+ = A (199)
(AT )+ = (A+ )T (200)
H + + H
(A ) = (A ) (201)
+ +
(A ) = (A ) (202)
(A+ A)AH = AH (203)
+ T T
(A A)A 6= A (204)
+ +
(cA) = (1/c)A (205)
+ T + T
A = (A A) A (206)
A+ = AT (AAT )+ (207)
(AT A)+ = A+ (AT )+ (208)
T + T + +
(AA ) = (A ) A (209)
+ H + H
A = (A A) A (210)
A+ = AH (AAH )+ (211)
H + + H +
(A A) = A (A ) (212)
H + H + +
(AA ) = (A ) A (213)
+ + + + +
(AB) = (A AB) (ABB ) (214)
f (AH A) f (0)I = A+ [f (AAH ) f (0)I]A (215)
H H +
f (AA ) f (0)I = A[f (A A) f (0)I]A (216)

where A Cnm .
Assume A to have full rank, then

(AA+ )(AA+ ) = AA+ (217)

+ + +
(A A)(A A) = A A (218)
Tr(AA+ ) = rank(AA+ ) (See [26]) (219)
+ +
Tr(A A) = rank(A A) (See [26]) (220)

For two matrices it hold that

(AB)+ = (A+ AB)+ (ABB+ )+ (221)

+ + +
(A B) = A B (222)

3.6.3 Construction
Assume that A has full rank, then
A nn Square rank(A) = n A+ = A1
A nm Broad rank(A) = n A+ = AT (AAT )1
A nm Tall rank(A) = m A+ = (AT A)1 AT
The so-called broad version is also known as right inverse and the tall ver-
sion as the left inverse.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 22
3.6 Pseudo Inverse 3 INVERSES

Assume A does not have full rank, i.e. A is n m and rank(A) = r <
min(n, m). The pseudo inverse A+ can be constructed from the singular value
decomposition A = UDVT , by

A+ = Vr D1 T
r Ur (223)

where Ur , Dr , and Vr are the matrices with the degenerated rows and columns
deleted. A different way is this: There do always exist two matrices C n r
and D r m of rank r, such that A = CD. Using these matrices it holds that

A+ = DT (DDT )1 (CT C)1 CT (224)

See [3].

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 23
 4 COMPLEX MATRICES

4 Complex Matrices
The complex scalar product r = pq can be written as
    
<r <p =p <q
= (225)
=r =p <p =q

4.1 Complex Derivatives

In order to differentiate an expression f (z) with respect to a complex z, the
Cauchy-Riemann equations have to be satisfied ([7]):
df (z) <(f (z)) =(f (z))
= +i (226)
dz <z <z
and
df (z) <(f (z)) =(f (z))
= i + (227)
dz =z =z
or in a more compact form:
f (z) f (z)
=i . (228)
=z <z
A complex function that satisfies the Cauchy-Riemann equations for points in a
region R is said yo be analytic in this region R. In general, expressions involving
complex conjugate or conjugate transpose do not satisfy the Cauchy-Riemann
equations. In order to avoid this problem, a more generalized definition of
complex derivative is used ([24], [6]):
Generalized Complex Derivative:
df (z) 1  f (z) f (z) 
= i . (229)
dz 2 <z =z

Conjugate Complex Derivative

df (z) 1  f (z) f (z) 

= +i . (230)
dz 2 <z =z

The Generalized Complex Derivative equals the normal derivative, when f is an

analytic function. For a non-analytic function such as f (z) = z , the derivative
equals zero. The Conjugate Complex Derivative equals zero, when f is an
analytic function. The Conjugate Complex Derivative has e.g been used by [21]
when deriving a complex gradient.
Notice:
df (z) f (z) f (z)
6= +i . (231)
dz <z =z
Complex Gradient Vector: If f is a real function of a complex vector z,
then the complex gradient vector is given by ([14, p. 798])
df (z)
f (z) = 2 (232)
dz
f (z) f (z)
= +i .
<z =z

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 24
4.1 Complex Derivatives 4 COMPLEX MATRICES

Complex Gradient Matrix: If f is a real function of a complex matrix Z,

then the complex gradient matrix is given by ([2])
df (Z)
f (Z) = 2 (233)
dZ
f (Z) f (Z)
= +i .
<Z =Z
These expressions can be used for gradient descent algorithms.

4.1.1 The Chain Rule for complex numbers

The chain rule is a little more complicated when the function of a complex
u = f (x) is non-analytic. For a non-analytic function, the following chain rule
can be applied ([7])
g(u) g u g u
= + (234)
x u x u x
g u  g  u
= +
u x u x
Notice, if the function is analytic, the second term reduces to zero, and the func-
tion is reduced to the normal well-known chain rule. For the matrix derivative
of a scalar function g(U), the chain rule can be written the following way:

g(U) Tr(( g(U) T

U ) U) Tr(( g(U) T
U ) U )
= + . (235)
X X X

4.1.2 Complex Derivatives of Traces

If the derivatives involve complex numbers, the conjugate transpose is often in-
volved. The most useful way to show complex derivative is to show the derivative
with respect to the real and the imaginary part separately. An easy example is:
Tr(X ) Tr(XH )
= = I (236)
<X <X

Tr(X ) Tr(XH )
i =i = I (237)
=X =X
Since the two results have the same sign, the conjugate complex derivative (230)
should be used.
Tr(X) Tr(XT )
= = I (238)
<X <X
Tr(X) Tr(XT )
i =i = I (239)
=X =X
Here, the two results have different signs, and the generalized complex derivative
(229) should be used. Hereby, it can be seen that (100) holds even if X is a
complex number.
Tr(AXH )
= A (240)
<X
Tr(AXH )
i = A (241)
=X

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 25
4.2 Higher order and non-linear derivatives 4 COMPLEX MATRICES

Tr(AX )
= AT (242)
<X
Tr(AX )
i = AT (243)
=X

Tr(XXH ) Tr(XH X)
= = 2<X (244)
<X <X
H
Tr(XX ) Tr(XH X)
i =i = i2=X (245)
=X =X
By inserting (244) and (245) in (229) and (230), it can be seen that

Tr(XXH )
= X (246)
X
Tr(XXH )
=X (247)
X

Since the function Tr(XXH ) is a real function of the complex matrix X, the
complex gradient matrix (233) is given by

Tr(XXH )
Tr(XXH ) = 2 = 2X (248)
X

4.1.3 Complex Derivative Involving Determinants

Here, a calculation example is provided. The objective is to find the derivative of
det(XH AX) with respect to X Cmn . The derivative is found with respect to
the real part and the imaginary part of X, by use of (42) and (37), det(XH AX)
can be calculated as (see App. B.1.4 for details)

det(XH AX) 1  det(XH AX) det(XH AX) 

= i
X 2 <X =X

T
= det(XH AX) (XH AX)1 XH A (249)

and the complex conjugate derivative yields

det(XH AX) 1  det(XH AX) det(XH AX) 

= +i
X 2 <X =X
= det(XH AX)AX(XH AX)1 (250)

4.2 Higher order and non-linear derivatives

(Ax)H (Ax) xH AH Ax
= (251)
x (Bx)H (Bx) x xH BH Bx
AH Ax xH AH AxBH Bx
= 2 H 2 (252)
x BBx (xH BH Bx)2

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 26
4.3 Inverse of complex sum 4 COMPLEX MATRICES

4.3 Inverse of complex sum

Given real matrices A, B find the inverse of the complex sum A + iB. Form
the auxiliary matrices

E = A + tB (253)
F = B tA, (254)

and find a value of t such that E1 exists. Then

(A + iB)1 = (1 it)(E + iF)1 (255)

1 1 1 1 1
= (1 it)((E + FE F) i(E + FE F) FE )(256)
1 1 1
= (1 it)(E + FE F) (I iFE ) (257)
= (E + FE1 F)1 ((I tFE1 ) i(tI + FE1 )) (258)
= (E + FE1 F)1 (I tFE1 )
i(E + FE1 F)1 (tI + FE1 ) (259)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 27
 5 SOLUTIONS AND DECOMPOSITIONS

5 Solutions and Decompositions

5.1 Solutions to linear equations
5.1.1 Simple Linear Regression
Assume we have data (xn , yn ) for n = 1, ..., N and are seeking the parameters
a, b R such that yi
= axi + b. With a least squares error function, the optimal
values for a, b can be expressed using the notation

x = (x1 , ..., xN )T y = (y1 , ..., yN )T 1 = (1, ..., 1)T RN 1

and

Rxx = xT x Rx1 = xT 1 R11 = 1T 1

Ryx = yT x Ry1 = yT 1

as    1  
a Rxx Rx1 Rx,y
= (260)
b Rx1 R11 Ry1

5.1.2 Existence in Linear Systems

Assume A is n m and consider the linear system

Ax = b (261)

Construct the augmented matrix B = [A b] then

Condition Solution
rank(A) = rank(B) = m Unique solution x
rank(A) = rank(B) < m Many solutions x
rank(A) < rank(B) No solutions x

5.1.3 Standard Square

Assume A is square and invertible, then

Ax = b x = A1 b (262)

5.1.4 Degenerated Square

Assume A is n n but of rank r < n. In that case, the system Ax = b is solved
by
x = A+ b
where A+ is the pseudo-inverse of the rank-deficient matrix, constructed as
described in section 3.6.3.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 28
5.1 Solutions to linear equations5 SOLUTIONS AND DECOMPOSITIONS

5.1.5 Cramers rule

The equation
Ax = b, (263)
where A is square has exactly one solution x if the ith element in x can be
found as
det B
xi = , (264)
det A
where B equals A, but the ith column in A has been substituted by b.

5.1.6 Over-determined Rectangular

Assume A to be n m, n > m (tall) and rank(A) = m, then
Ax = b x = (AT A)1 AT b = A+ b (265)
that is if there exists a solution x at all! If there is no solution the following
can be useful:
Ax = b xmin = A+ b (266)
Now xmin is the vector x which minimizes ||Ax b||2 , i.e. the vector which is
least wrong. The matrix A+ is the pseudo-inverse of A. See [3].

5.1.7 Under-determined Rectangular

Assume A is n m and n < m (broad) and rank(A) = n.
Ax = b xmin = AT (AAT )1 b (267)
The equation have many solutions x. But xmin is the solution which minimizes
||Ax b||2 and also the solution with the smallest norm ||x||2 . The same holds
for a matrix version: Assume A is n m, X is m n and B is n n, then
AX = B Xmin = A+ B (268)
The equation have many solutions X. But Xmin is the solution which minimizes
||AX B||2 and also the solution with the smallest norm ||X||2 . See [3].
Similar but different: Assume A is square n n and the matrices B0 , B1
are n N , where N > n, then if B0 has maximal rank
AB0 = B1 Amin = B1 BT0 (B0 BT0 )1 (269)
where Amin denotes the matrix which is optimal in a least square sense. An
interpretation is that A is the linear approximation which maps the columns
vectors of B0 into the columns vectors of B1 .

5.1.8 Linear form and zeros

Ax = 0, x A=0 (270)

5.1.9 Square form and zeros

If A is symmetric, then
xT Ax = 0, x A=0 (271)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 29
5.2 Eigenvalues and Eigenvectors5 SOLUTIONS AND DECOMPOSITIONS

5.1.10 The Lyapunov Equation

AX + XB = C (272)
vec(X) = (I A + BT I)1 vec(C) (273)

Sec 10.2.1 and 10.2.2 for details on the Kronecker product and the vec op-
erator.

5.1.11 Encapsulating Sum

P
n An XBn = C (274)
P T
1
vec(X) = n Bn An vec(C) (275)

See Sec 10.2.1 and 10.2.2 for details on the Kronecker product and the vec
operator.

5.2 Eigenvalues and Eigenvectors

5.2.1 Definition
The eigenvectors vi and eigenvalues i are the ones satisfying

Avi = i vi (276)

5.2.2 Decompositions
For matrices A with as many distinct eigenvalues as dimensions, the following
holds, where the columns of V are the eigenvectors and (D)ij = ij i ,

AV = VD (277)

For defective matrices A, which is matrices which has fewer distinct eigenvalues
than dimensions, the following decomposition called Jordan canonical form,
holds
AV = VJ (278)
where J is a block diagonal matrix with the blocks Ji = i I + N. The matrices
Ji have dimensionality as the number of identical eigenvalues equal to i , and N
is square matrix of same size with 1 on the super diagonal and zero elsewhere.
It also holds that for all matrices A there exists matrices V and R such that

AV = VR (279)

where R is upper triangular with the eigenvalues i on its diagonal.

5.2.3 General Properties

Assume that A Rnm and B Rmn ,

eig(AB) = eig(BA) (280)

rank(A) = r At most r non-zero i (281)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 30
5.3 Singular Value Decomposition
5 SOLUTIONS AND DECOMPOSITIONS

5.2.4 Symmetric
Assume A is symmetric, then
VVT = I (i.e. V is orthogonal) (282)
i R (i.e. i is real) (283)
p P p
Tr(A ) = i i (284)
eig(I + cA) = 1 + ci (285)
eig(A cI) = i c (286)
eig(A 1
) = 1
i (287)
For a symmetric, positive matrix A,
eig(AT A) = eig(AAT ) = eig(A) eig(A) (288)

5.2.5 Characteristic polynomial

The characteristic polynomial for the matrix A is
0 = det(A I) (289)
n n1 n2 n
= g1 + g2 ... + (1) gn (290)
Note that the coefficients gj for j = 1, ..., n are the n invariants under rotation
of A. Thus, gj is the sum of the determinants of all the sub-matrices of A taken
j rows and columns at a time. That is, g1 is the trace of A, and g2 is the sum
of the determinants of the n(n 1)/2 sub-matrices that can be formed from A
by deleting all but two rows and columns, and so on see [17].

5.3 Singular Value Decomposition

Any n m matrix A can be written as
A = UDVT , (291)
where
U eigenvectors of AAT
= p nn
D = diag(eig(AAT )) nm (292)
V = eigenvectors of AT A mm

5.3.1 Symmetric Square decomposed into squares

Assume A to be n n and symmetric. Then
     T 
A = V D V , (293)
where D is diagonal with the eigenvalues of A, and V is orthogonal and the
eigenvectors of A.

5.3.2 Square decomposed into squares

Assume A Rnn . Then
UT
     
A = V D , (294)
where D is diagonal with the square root of the eigenvalues of AAT , V is the
eigenvectors of AAT and UT is the eigenvectors of AT A.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 31
5.4 Triangular Decomposition 5 SOLUTIONS AND DECOMPOSITIONS

5.3.3 Square decomposed into rectangular

Assume V D UT = 0 then we can expand the SVD of A into
  T 
    D 0 U
A = V V , (295)
0 D UT

where the SVD of A is A = VDUT .

5.3.4 Rectangular decomposition I

Assume A is n m, V is n n, D is n n, UT is n m

UT
     
A = V D , (296)

where D is diagonal with the square root of the eigenvalues of AAT , V is the
eigenvectors of AAT and UT is the eigenvectors of AT A.

5.3.5 Rectangular decomposition II

Assume A is n m, V is n m, D is m m, UT is m m

UT
   
A = V D (297)

5.3.6 Rectangular decomposition III

Assume A is n m, V is n n, D is n m, UT is m m

UT
    
A = V D , (298)

where D is diagonal with the square root of the eigenvalues of AAT , V is the
eigenvectors of AAT and UT is the eigenvectors of AT A.

5.4 Triangular Decomposition

5.5 LU decomposition
Assume A is a square matrix with non-zero leading principal minors, then

A = LU (299)

where L is a unique unit lower triangular matrix and U is a unique upper

triangular matrix.

5.5.1 Cholesky-decomposition
Assume A is a symmetric positive definite square matrix, then

A = UT U = LLT , (300)

where U is a unique upper triangular matrix and L is a lower triangular matrix.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 32
5.6 LDM decomposition 5 SOLUTIONS AND DECOMPOSITIONS

5.6 LDM decomposition

Assume A is a square matrix with non-zero leading principal minors1 , then

A = LDMT (301)

where L, M are unique unit lower triangular matrices and D is a unique diagonal
matrix.

5.7 LDL decompositions

The LDL decomposition are special cases of the LDM decomposition. Assume
A is a non-singular symmetric definite square matrix, then

A = LDLT = LT DL (302)

where L is a unit lower triangular matrix and D is a diagonal matrix. If A is

also positive definite, then D has strictly positive diagonal entries.
1 If the matrix that corresponds to a principal minor is a quadratic upper-left part of the

larger matrix (i.e., it consists of matrix elements in rows and columns from 1 to k), then the
principal minor is called a leading principal minor. For an n times n square matrix, there are
n leading principal minors. [31]

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 33
 6 STATISTICS AND PROBABILITY

6 Statistics and Probability

6.1 Definition of Moments
Assume x Rn1 is a random variable

6.1.1 Mean
The vector of means, m, is defined by

(m)i = hxi i (303)

6.1.2 Covariance
The matrix of covariance M is defined by

(M)ij = h(xi hxi i)(xj hxj i)i (304)

or alternatively as
M = h(x m)(x m)T i (305)

6.1.3 Third moments

The matrix of third centralized moments in some contexts referred to as
coskewness is defined using the notation
(3)
mijk = h(xi hxi i)(xj hxj i)(xk hxk i)i (306)

as h i
(3) (3)
M3 = m::1 m::2 ...m(3)
::n (307)

where : denotes all elements within the given index. M3 can alternatively be
expressed as
M3 = h(x m)(x m)T (x m)T i (308)

6.1.4 Fourth moments

The matrix of fourth centralized moments in some contexts referred to as
cokurtosis is defined using the notation
(4)
mijkl = h(xi hxi i)(xj hxj i)(xk hxk i)(xl hxl i)i (309)

as
h i
(4) (4) (4) (4) (4) (4) (4) (4)
M4 = m::11 m::21 ...m::n1 |m::12 m::22 ...m::n2 |...|m::1n m::2n ...m(4)
::nn (310)

or alternatively as

M4 = h(x m)(x m)T (x m)T (x m)T i (311)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 34
6.2 Expectation of Linear Combinations
6 STATISTICS AND PROBABILITY

6.2 Expectation of Linear Combinations

6.2.1 Linear Forms
Assume X and x to be a matrix and a vector of random variables. Then (see
See [26])

E[AXB + C] = AE[X]B + C (312)

T
Var[Ax] = AVar[x]A (313)
T
Cov[Ax, By] = ACov[x, y]B (314)

Assume x to be a stochastic vector with mean m, then (see [7])

E[Ax + b] = Am + b (315)
E[Ax] = Am (316)
E[x + b] = m+b (317)

6.2.2 Quadratic Forms

Assume A is symmetric, c = E[x] and = Var[x]. Assume also that all
coordinates xi are independent, have the same central moments 1 , 2 , 3 , 4
and denote a = diag(A). Then (See [26])

E[xT Ax] = Tr(A) + cT Ac (318)

T
Var[x Ax] = 222 Tr(A2 ) + 42 cT A2 c + 43 cT Aa + (4 322 )aT a (319)

Also, assume x to be a stochastic vector with mean m, and covariance M. Then

(see [7])

E[(Ax + a)(Bx + b)T ] = AMBT + (Am + a)(Bm + b)T (320)

T T
E[xx ] = M + mm (321)
E[xaT x] = (M + mmT )a (322)
E[xT axT ] = aT (M + mmT ) (323)
T T T
E[(Ax)(Ax) ] = A(M + mm )A (324)
T T
E[(x + a)(x + a) ] = M + (m + a)(m + a) (325)

E[(Ax + a)T (Bx + b)] = Tr(AMBT ) + (Am + a)T (Bm + b) (326)

E[xT x] = Tr(M) + mT m (327)
T
E[x Ax] = Tr(AM) + mT Am (328)
T T T
E[(Ax) (Ax)] = Tr(AMA ) + (Am) (Am) (329)
T T
E[(x + a) (x + a)] = Tr(M) + (m + a) (m + a) (330)

See [7].

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 35
6.3 Weighted Scalar Variable 6 STATISTICS AND PROBABILITY

6.2.3 Cubic Forms

Assume x to be a stochastic vector with independent coordinates, mean m,
covariance M and central moments v3 = E[(x m)3 ]. Then (see [7])

E[(Ax + a)(Bx + b)T (Cx + c)] = Adiag(BT C)v3

+Tr(BMCT )(Am + a)
+AMCT (Bm + b)
+(AMBT + (Am + a)(Bm + b)T )(Cm + c)
T
E[xx x] = v3 + 2Mm + (Tr(M) + mT m)m
E[(Ax + a)(Ax + a)T (Ax + a)] = Adiag(AT A)v3
+[2AMAT + (Ax + a)(Ax + a)T ](Am + a)
+Tr(AMAT )(Am + a)
T T
E[(Ax + a)b (Cx + c)(Dx + d) ] = (Ax + a)bT (CMDT + (Cm + c)(Dm + d)T )
+(AMCT + (Am + a)(Cm + c)T )b(Dm + d)T
+bT (Cm + c)(AMDT (Am + a)(Dm + d)T )

6.3 Weighted Scalar Variable

Assume x Rn1 is a random variable, w Rn1 is a vector of constants and
y is the linear combination y = wT x. Assume further that m, M2 , M3 , M4
denotes the mean, covariance, and central third and fourth moment matrix of
the variable x. Then it holds that

hyi = wT m (331)
2 T
h(y hyi) i = w M2 w (332)
h(y hyi)3 i = wT M3 w w (333)
h(y hyi)4 i = wT M4 w w w (334)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 36
 7 MULTIVARIATE DISTRIBUTIONS

7 Multivariate Distributions
7.1 Cauchy
The density function for a Cauchy distributed vector t RP 1 , is given by

( 1+P
2 ) det()1/2
p(t|, ) = P/2 (335)
(1/2) 1 + (t )T 1 (t )(1+P )/2


where is the location, is positive definite, and denotes the gamma func-
tion. The Cauchy distribution is a special case of the Student-t distribution.

7.2 Dirichlet
The Dirichlet distribution is a kind of inverse distribution compared to the
multinomial distribution on the bounded continuous variate x = [x1 , . . . , xP ]
[16, p. 44] P 
P P
p p Y 1
p(x|) = QP xp p
p ( p ) p

7.3 Normal
The normal distribution is also known as a Gaussian distribution. See sec. 8.

7.4 Normal-Inverse Gamma

7.5 Gaussian
See sec. 8.

7.6 Multinomial
If the vector n contains counts, i.e. (n)i 0, 1, 2, ..., then the discrete multino-
mial disitrbution for n is given by
d d
n! Y X
P (n|a, n) = ani , ni = n (336)
n1 ! . . . nd ! i i i
P
where ai are probabilities, i.e. 0 ai 1 and i ai = 1.

7.7 Students t
The density of a Student-t distributed vector t RP 1 , is given by

( +P
2 ) det()1/2
p(t|, , ) = ()P/2 (337)
(/2) 1 + 1 (t )T 1 (t )(+P )/2


where is the location, the scale matrix is symmetric, positive definite,

is the degrees of freedom, and denotes the gamma function. For = 1, the
Student-t distribution becomes the Cauchy distribution (see sec 7.1).

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 37
7.8 Wishart 7 MULTIVARIATE DISTRIBUTIONS

7.7.1 Mean
E(t) = , >1 (338)

7.7.2 Variance

cov(t) = , >2 (339)
2

7.7.3 Mode
The notion mode meaning the position of the most probable value

mode(t) = (340)

7.7.4 Full Matrix Version

If instead of a vector t RP 1 one has a matrix T RP N , then the Student-t
distribution for T is
P
Y [( + P p + 1)/2]
p(T|M, , , ) = N P/2
p=1
[( p + 1)/2]

det()/2 det()N/2
(+P )/2
det 1 + (T M)1 (T M)T

(341)

where M is the location, is the rescaling matrix, is positive definite, is

the degrees of freedom, and denotes the gamma function.

7.8 Wishart
The central Wishart distribution for M RP P , M is positive definite, where
m can be regarded as a degree of freedom parameter [16, equation 3.8.1] [8,
section 2.5],[11]
1
p(M|, m) = QP
2mP/2 P (P 1)/4 p [ 12 (m + 1 p)]
det()m/2 det(M)(mP 1)/2
 
1 1
exp Tr( M) (342)
2

7.8.1 Mean
E(M) = m (343)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 38
7.9 Wishart, Inverse 7 MULTIVARIATE DISTRIBUTIONS

7.9 Wishart, Inverse

The (normal) Inverse Wishart distribution for M RP P , M is positive defi-
nite, where m can be regarded as a degree of freedom parameter [11]
1
p(M|, m) = QP
2mP/2 P (P 1)/4 p [ 12 (m + 1 p)]
det()m/2 det(M)(mP 1)/2
 
1
exp Tr(M1 ) (344)
2

7.9.1 Mean
1
E(M) = (345)
mP 1

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 39
 8 GAUSSIANS

8 Gaussians
8.1 Basics
8.1.1 Density and normalization
The density of x N (m, ) is
 
1 1 T 1
p(x) = p exp (x m) (x m) (346)
det(2) 2

Note that if x is d-dimensional, then det(2) = (2)d det().

Integration and normalization
Z  
1 p
exp (x m)T 1 (x m) dx = det(2)
2
Z    
1 p 1
exp xT 1 x + mT 1 x dx = det(2) exp mT 1 m
2 2
Z    
1 T T
p
1
1 T T
exp x Ax + c x dx = det(2A ) exp c A c
2 2

If X = [x1 x2 ...xn ] and C = [c1 c2 ...cn ], then

Z    
1 T T
p
1
n 1 T 1
exp Tr(X AX) + Tr(C X) dX = det(2A ) exp Tr(C A C)
2 2

The derivatives of the density are

p(x)
= p(x)1 (x m) (347)
x
2p  
= p(x) 1 (x m)(x m)T 1 1 (348)
xxT

8.1.2 Marginal Distribution

Assume x Nx (, ) where
     
xa a a c
x= = = (349)
xb b Tc b

then

p(xa ) = Nxa (a , a ) (350)

p(xb ) = Nxb (b , b ) (351)

8.1.3 Conditional Distribution

Assume x Nx (, ) where
     
xa a a c
x= = = (352)
xb b Tc b

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 40
8.1 Basics 8 GAUSSIANS

then
n
a = a + c 1
b (xb b ) (353)
p(xa |xb ) = Nxa (a , a )
a = a c 1
b c
T

n =
b b + Tc 1
a (xa a )
p(xb |xa ) = Nxb (b , b ) (354)
b = b Tc 1
a c

Note, that the covariance matrices are the Schur complement of the block ma-
trix, see 9.1.5 for details.

8.1.4 Linear combination

Assume x N (mx , x ) and y N (my , y ) then

Ax + By + c N (Amx + Bmy + c, Ax AT + By BT ) (355)

8.1.5 Rearranging Means

p
det(2(AT 1 A)1 )
NAx [m, ] = p Nx [A1 m, (AT 1 A)1 ] (356)
det(2)
If A is square and invertible, it simplifies to
1
NAx [m, ] = Nx [A1 m, (AT 1 A)1 ] (357)
| det(A)|

8.1.6 Rearranging into squared form

If A is symmetric, then
1 1 1
= (x A1 b)T A(x A1 b) + bT A1 b
xT Ax + bT x
2 2 2
1 1 1
Tr(XT AX) + Tr(BT X) = Tr[(X A1 B)T A(X A1 B)] + Tr(BT A1 B)
2 2 2

8.1.7 Sum of two squared forms

In vector formulation (assuming 1 , 2 are symmetric)
1
(x m1 )T 1
1 (x m1 ) (358)
2
1
(x m2 )T 1
2 (x m2 ) (359)
2
1
= (x mc )T 1
c (x mc ) + C (360)
2

1
c = 1 1
1 + 2 (361)
mc = (1
1 + 1
2 )
1
(1
1 m1
+ 1
2 m2 ) (362)
1 T 1
C = (m + mT2 1 1 1 1
2 )(1 + 2 ) (1 1
1 m1 + 2 m2 )(363)
2 1 1
1 
mT1 1 T 1
1 m1 + m2 2 m2 (364)
2

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 41
8.2 Moments 8 GAUSSIANS

In a trace formulation (assuming 1 , 2 are symmetric)

1
Tr((X M1 )T 1
1 (X M1 )) (365)
2
1
Tr((X M2 )T 1
2 (X M2 )) (366)
2
1
= Tr[(X Mc )T 1
c (X Mc )] + C (367)
2

1
c = 1
1 + 2
1
(368)
1 1 1 1 1
Mc = (1 + 2 ) (1 M1 + 2 M2 ) (369)
1 h 1 i
C = Tr (1 M1 + 1
2 M 2 )T
(1
1 + 1 1
2 ) (1
1 M1 + 1
2 M2 )
2
1
Tr(MT1 1 T 1
1 M1 + M2 2 M2 ) (370)
2

8.1.8 Product of gaussian densities

Let Nx (m, ) denote a density of x, then

Nx (m1 , 1 ) Nx (m2 , 2 ) = cc Nx (mc , c ) (371)

cc = Nm1 (m2 , (1 + 2 ))
 
1 1
= p exp (m1 m2 )T (1 + 2 )1 (m1 m2 )
det(2(1 + 2 )) 2
mc = (1 1 1
1 + 2 ) (1 1
1 m1 + 2 m2 )
c = (1 1 1
1 + 2 )

but note that the product is not normalized as a density of x.

8.2 Moments
8.2.1 Mean and covariance of linear forms
First and second moments. Assume x N (m, )

E(x) = m (372)

Cov(x, x) = Var(x) = = E(xxT ) E(x)E(xT ) = E(xxT ) mmT (373)

As for any other distribution is holds for gaussians that

E[Ax] = AE[x] (374)

T
Var[Ax] = AVar[x]A (375)
T
Cov[Ax, By] = ACov[x, y]B (376)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 42
8.2 Moments 8 GAUSSIANS

8.2.2 Mean and variance of square forms

Mean and variance of square forms: Assume x N (m, )

E(xxT ) = + mmT (377)

T T
E[x Ax] = Tr(A) + m Am (378)
T T
Var(x Ax) = Tr[A(A + A )] + ...
+mT (A + AT )(A + AT )m (379)
0 T 0
E[(x m ) A(x m )] = (m m0 )T A(m m0 ) + Tr(A) (380)

If = 2 I and A is symmetric, then

Var(xT Ax) = 2 4 Tr(A2 ) + 4 2 mT A2 m (381)

Assume x N (0, 2 I) and A and B to be symmetric, then

Cov(xT Ax, xT Bx) = 2 4 Tr(AB) (382)

8.2.3 Cubic forms

Assume x to be a stochastic vector with independent coordinates, mean m and
covariance M

E[xbT xxT ] = mbT (M + mmT ) + (M + mmT )bmT

+bT m(M mmT ) (383)

8.2.4 Mean of Quartic Forms

E[xxT xxT ] = 2( + mmT )2 + mT m( mmT )

+Tr()( + mmT )
T T
E[xx Axx ] = ( + mmT )(A + AT )( + mmT )
+mT Am( mmT ) + Tr[A]( + mmT )
E[xT xxT x] = 2Tr(2 ) + 4mT m + (Tr() + mT m)2
E[xT AxxT Bx] = Tr[A(B + BT )] + mT (A + AT )(B + BT )m
+(Tr(A) + mT Am)(Tr(B) + mT Bm)

E[aT xbT xcT xdT x]

= (aT ( + mmT )b)(cT ( + mmT )d)
+(aT ( + mmT )c)(bT ( + mmT )d)
+(aT ( + mmT )d)(bT ( + mmT )c) 2aT mbT mcT mdT m

E[(Ax + a)(Bx + b)T (Cx + c)(Dx + d)T ]

= [ABT + (Am + a)(Bm + b)T ][CDT + (Cm + c)(Dm + d)T ]
+[ACT + (Am + a)(Cm + c)T ][BDT + (Bm + b)(Dm + d)T ]
+(Bm + b)T (Cm + c)[ADT (Am + a)(Dm + d)T ]
+Tr(BCT )[ADT + (Am + a)(Dm + d)T ]

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 43
8.3 Miscellaneous 8 GAUSSIANS

E[(Ax + a)T (Bx + b)(Cx + c)T (Dx + d)]

= Tr[A(CT D + DT C)BT ]
+[(Am + a)T B + (Bm + b)T A][CT (Dm + d) + DT (Cm + c)]
+[Tr(ABT ) + (Am + a)T (Bm + b)][Tr(CDT ) + (Cm + c)T (Dm + d)]
See [7].

8.2.5 Moments
X
E[x] = k mk (384)
k
XX
Cov(x) = k k0 (k + mk mTk mk mTk0 ) (385)
k k0

8.3 Miscellaneous
8.3.1 Whitening
Assume x N (m, ) then
z = 1/2 (x m) N (0, I) (386)
Conversely having z N (0, I) one can generate data x N (m, ) by setting
x = 1/2 z + m N (m, ) (387)
Note that 1/2 means the matrix which fulfils 1/2 1/2 = , and that it exists
and is unique since is positive definite.

8.3.2 The Chi-Square connection

Assume x N (m, ) and x to be n dimensional, then
z = (x m)T 1 (x m) 2n (388)
where 2n denotes the Chi square distribution with n degrees of freedom.

8.3.3 Entropy
Entropy of a D-dimensional gaussian
Z p D
H(x) = N (m, ) ln N (m, )dx = ln det(2) + (389)
2

8.4 Mixture of Gaussians

8.4.1 Density
The variable x is distributed as a mixture of gaussians if it has the density
K  
X 1 1 T 1
p(x) = k p exp (x mk ) k (x mk ) (390)
det(2k ) 2
k=1

where k sum to 1 and the k all are positive definite.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 44
8.4 Mixture of Gaussians 8 GAUSSIANS

8.4.2 Derivatives
P
Defining p(s) = k k Ns (k , k ) one get

ln p(s) j Ns (j , j )
= P ln[j Ns (j , j )] (391)
j k k N s (k , k ) j
j Ns (j , j ) 1
= P (392)
k k Ns (k , k ) j
ln p(s) j Ns (j , j )
= P ln[j Ns (j , j )] (393)
j
k k s N (k , k ) j
j Ns (j , j )  1 
= P j (s j ) (394)
k k Ns (k , k )
ln p(s) j Ns (j , j )
= P ln[j Ns (j , j )] (395)
j
k k s N (k , k ) j
j Ns (j , j ) 1 
1 1 T 1

= P j + j (s j )(s j ) j (396)

k k s N (k , k ) 2

But k and k needs to be constrained.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 45
 9 SPECIAL MATRICES

9 Special Matrices
9.1 Block matrices
Let Aij denote the ijth block of A.

9.1.1 Multiplication
Assuming the dimensions of the blocks matches we have
    
A11 A12 B11 B12 A11 B11 + A12 B21 A11 B12 + A12 B22
=
A21 A22 B21 B22 A21 B11 + A22 B21 A21 B12 + A22 B22

9.1.2 The Determinant

The determinant can be expressed as by the use of

C1 = A11 A12 A1
22 A21 (397)
C2 = A22 A21 A1
11 A12 (398)

as  
A11 A12
det = det(A22 ) det(C1 ) = det(A11 ) det(C2 )
A21 A22

9.1.3 The Inverse

The inverse can be expressed as by the use of

C1 = A11 A12 A1
22 A21 (399)
C2 = A22 A21 A1
11 A12 (400)

as 1
C1 A1 1
  
A11 A12 1 11 A12 C2
=
A21 A22 C2 A21 A1
1
11 C21

A1 1 1 1
C1 1
 
11 + A11 A12 C2 A21 A11 1 A12 A22
= 1 1
A22 A21 C1 A22 + A22 A21 C1 A12 A1
1 1 1
22

9.1.4 Block diagonal

For block diagonal matrices we have
1
(A11 )1
  
A11 0 0
= (401)
0 A22 0 (A22 )1
 
A11 0
det = det(A11 ) det(A22 ) (402)
0 A22

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 46
9.2 Discrete Fourier Transform Matrix, The 9 SPECIAL MATRICES

9.1.5 Schur complement

Regard the matrix  
A11 A12
A21 A22
The Schur complement of block A11 of the matrix above is the matrix (denoted
C2 in the text above)
A22 A21 A1
11 A12

The Schur complement of block A22 of the matrix above is the matrix (denoted
C1 in the text above)
A11 A12 A1
22 A21

Using the Schur complement, one can rewrite the inverse of a block matrix
 1
A11 A12
A21 A22
(A11 A12 A1 1
A12 A1
   
I 0 22 A21 ) 0 I 22
=
A1
22 A21 I 0 A1
22 0 I

The Schur complement is useful when solving linear systems of the form
    
A11 A12 x1 b1
=
A21 A22 x2 b2

which has the following equation for x1

(A11 A12 A1 1
22 A21 )x1 = b1 A12 A22 b2

When the appropriate inverses exists, this can be solved for x1 which can then
be inserted in the equation for x2 to solve for x2 .

9.2 Discrete Fourier Transform Matrix, The

The DFT matrix is an N N symmetric matrix WN , where the k, nth element
is given by
j2kn
WNkn = e N (403)
Thus the discrete Fourier transform (DFT) can be expressed as
N
X 1
X(k) = x(n)WNkn . (404)
n=0

Likewise the inverse discrete Fourier transform (IDFT) can be expressed as

N 1
1 X
x(n) = X(k)WNkn . (405)
N
k=0

The DFT of the vector x = [x(0), x(1), , x(N 1)]T can be written in matrix
form as
X = WN x, (406)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 47
9.3 Hermitian Matrices and skew-Hermitian 9 SPECIAL MATRICES

where X = [X(0), X(1), , x(N 1)]T . The IDFT is similarly given as

x = W1
N X. (407)

Some properties of WN exist:

1
W1 = W (408)
N
N N
WN WN = NI (409)
WN = WH
N (410)
j2
If WN = e N , then [23]
m+N/2
WN = WNm (411)
Notice, the DFT matrix is a Vandermonde Matrix.
The following important relation between the circulant matrix and the dis-
crete Fourier transform (DFT) exists

TC = W1
N (I (WN t))WN , (412)

where t = [t0 , t1 , , tn1 ]T is the first row of TC .

9.3 Hermitian Matrices and skew-Hermitian

A matrix A Cmn is called Hermitian if

AH = A

For real valued matrices, Hermitian and symmetric matrices are equivalent.

A is Hermitian xH Ax R, x Cn1 (413)

A is Hermitian eig(A) R (414)

Note that
A = B + iC
where B, C are hermitian, then

A + AH A AH
B= , C=
2 2i

9.3.1 Skew-Hermitian
A matrix A is called skew-hermitian if

A = AH

For real valued matrices, skew-Hermitian and skew-symmetric matrices are

equivalent.

A Hermitian iA is skew-hermitian (415)

H H H
A skew-Hermitian x Ay = x A y, x, y (416)
A skew-Hermitian eig(A) = i, R (417)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 48
9.4 Idempotent Matrices 9 SPECIAL MATRICES

9.4 Idempotent Matrices

A matrix A is idempotent if
AA = A
Idempotent matrices A and B, have the following properties

An = A, forn = 1, 2, 3, ... (418)

IA is idempotent (419)
AH is idempotent (420)
H
IA is idempotent (421)
If AB = BA AB is idempotent (422)
rank(A) = Tr(A) (423)
A(I A) = 0 (424)
(I A)A = 0 (425)
+
A = A (426)
f (sI + tA) = (I A)f (s) + Af (s + t) (427)

Note that A I is not necessarily idempotent.

9.4.1 Nilpotent
A matrix A is nilpotent if
A2 = 0
A nilpotent matrix has the following property:

f (sI + tA) = If (s) + tAf 0 (s) (428)

9.4.2 Unipotent
A matrix A is unipotent if
AA = I
A unipotent matrix has the following property:

f (sI + tA) = [(I + A)f (s + t) + (I A)f (s t)]/2 (429)

9.5 Orthogonal matrices

If a square matrix Q is orthogonal, if and only if,

QT Q = QQT = I

and then Q has the following properties

Its eigenvalues are placed on the unit circle.

Its eigenvectors are unitary, i.e. have length one.

The inverse of an orthogonal matrix is orthogonal too.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 49
9.6 Positive Definite and Semi-definite Matrices 9 SPECIAL MATRICES

Basic properties for the orthogonal matrix Q

Q1 = QT
T
Q = Q
T
QQ = I
T
Q Q = I
det(Q) = 1

9.5.1 Ortho-Sym
A matrix Q+ which simultaneously is orthogonal and symmetric is called an
ortho-sym matrix [20]. Hereby
QT+ Q+ = I (430)
Q+ = QT+ (431)
The powers of an ortho-sym matrix are given by the following rule
1 + (1)k 1 + (1)k+1
Qk+ = I+ Q+ (432)
2 2
1 + cos(k) 1 cos(k)
= I+ Q+ (433)
2 2

9.5.2 Ortho-Skew
A matrix which simultaneously is orthogonal and antisymmetric is called an
ortho-skew matrix [20]. Hereby
QH
Q = I (434)
Q = QH
(435)
The powers of an ortho-skew matrix are given by the following rule
ik + (i)k ik (i)k
Qk = Ii Q (436)
2 2

= cos(k )I + sin(k )Q (437)
2 2

9.5.3 Decomposition
A square matrix A can always be written as a sum of a symmetric A+ and an
antisymmetric matrix A
A = A+ + A (438)

9.6 Positive Definite and Semi-definite Matrices

9.6.1 Definitions
A matrix A is positive definite if and only if
xT Ax > 0, x 6= 0 (439)
A matrix A is positive semi-definite if and only if
xT Ax 0, x (440)
Note that if A is positive definite, then A is also positive semi-definite.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 50
9.6 Positive Definite and Semi-definite Matrices 9 SPECIAL MATRICES

9.6.2 Eigenvalues
The following holds with respect to the eigenvalues:
H
A pos. def. eig( A+A
2 )>0
H (441)
A pos. semi-def. eig( A+A
2 )0

9.6.3 Trace
The following holds with respect to the trace:

A pos. def. Tr(A) > 0

(442)
A pos. semi-def. Tr(A) 0

9.6.4 Inverse
If A is positive definite, then A is invertible and A1 is also positive definite.

9.6.5 Diagonal
If A is positive definite, then Aii > 0, i

9.6.6 Decomposition I
The matrix A is positive semi-definite of rank r there exists a matrix B of
rank r such that A = BBT

The matrix A is positive definite there exists an invertible matrix B such

that A = BBT

9.6.7 Decomposition II
Assume A is an n n positive semi-definite, then there exists an n r matrix
B of rank r such that BT AB = I.

9.6.8 Equation with zeros

Assume A is positive semi-definite, then XT AX = 0 AX = 0

9.6.9 Rank of product

Assume A is positive definite, then rank(BABT ) = rank(B)

9.6.10 Positive definite property

If A is n n positive definite and B is r n of rank r, then BABT is positive
definite.

9.6.11 Outer Product

If X is n r, where n r and rank(X) = n, then XXT is positive definite.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 51
9.7 Singleentry Matrix, The 9 SPECIAL MATRICES

9.6.12 Small pertubations

If A is positive definite and B is symmetric, then A tB is positive definite for
sufficiently small t.

9.6.13 Hadamard inequality

If A is a positive definite or semi-definite matrix, then
Y
det(A) Aii
i

See [15, pp.477]

9.6.14 Hadamard product relation

Assume that P = AAT and Q = BBT are semi positive definite matrices, it
then holds that
P Q = RRT
where the columns of R are constructed as follows: ri+(j1)NA = ai bj , for
i = 1, 2, ..., NA and j = 1, 2, ..., NB . The result is unpublished, but reported by
Pavel Sakov and Craig Bishop.

9.7 Singleentry Matrix, The

9.7.1 Definition
The single-entry matrix Jij Rnn is defined as the matrix which is zero
everywhere except in the entry (i, j) in which it is 1. In a 4 4 example one
might have
0 0 0 0
0 0 1 0
J23 =
0 0 0 0
(443)
0 0 0 0
The single-entry matrix is very useful when working with derivatives of expres-
sions involving matrices.

9.7.2 Swap and Zeros

Assume A to be n m and Jij to be m p
AJij = 0 0 . . . Ai
 
... 0 (444)
i.e. an n p matrix of zeros with the i.th column of A in place of the j.th
column. Assume A to be n m and Jij to be p n

0
..
.

0
Jij A =

Aj (445)

0

.
..
0

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 52
9.7 Singleentry Matrix, The 9 SPECIAL MATRICES

i.e. an p m matrix of zeros with the j.th row of A in the placed of the i.th
row.

9.7.3 Rewriting product of elements

Aki Bjl = (Aei eTj B)kl = (AJij B)kl (446)

Aik Blj = (A ei eTj BT )kl
T
= T
(A J B )kl ij T
(447)
Aik Bjl = (AT ei eTj B)kl = T
(A J B)kl ij
(448)
Aki Blj = (Aei eTj BT )kl = ij
(AJ B )kl T
(449)

9.7.4 Properties of the Singleentry Matrix

If i = j
Jij Jij = Jij (Jij )T (Jij )T = Jij
Jij (Jij )T = Jij (Jij )T Jij = Jij
If i 6= j
Jij Jij = 0 (Jij )T (Jij )T = 0
Jij (Jij )T = Jii (Jij )T Jij = Jjj

9.7.5 The Singleentry Matrix in Scalar Expressions

Assume A is n m and J is m n, then

Tr(AJij ) = Tr(Jij A) = (AT )ij (450)

Assume A is n n, J is n m and B is m n, then

Tr(AJij B) = (AT BT )ij (451)

Tr(AJji B) = (BA)ij (452)
Tr(AJij Jij B) = diag(AT BT )ij (453)

Assume A is n n, Jij is n m B is m n, then

xT AJij Bx = (AT xxT BT )ij (454)

T ij ij
x AJ J Bx = diag(AT xxT BT )ij (455)

9.7.6 Structure Matrices

The structure matrix is defined by
A
= Sij (456)
Aij

If A has no special structure then

Sij = Jij (457)

If A is symmetric then
Sij = Jij + Jji Jij Jij (458)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 53
9.8 Symmetric, Skew-symmetric/Antisymmetric 9 SPECIAL MATRICES

9.8 Symmetric, Skew-symmetric/Antisymmetric

9.8.1 Symmetric
The matrix A is said to be symmetric if

A = AT (459)

Symmetric matrices have many important properties, e.g. that their eigenvalues
are real and eigenvectors orthogonal.

9.8.2 Skew-symmetric/Antisymmetric
The antisymmetric matrix is also known as the skew symmetric matrix. It has
the following property from which it is defined

A = AT (460)

Hereby, it can be seen that the antisymmetric matrices always have a zero
diagonal. The n n antisymmetric matrices also have the following properties.

det(AT ) = det(A) = (1)n det(A) (461)

det(A) = det(A) = 0, if n is odd (462)

The eigenvalues of an antisymmetric matrix are placed on the imaginary axis

and the eigenvectors are unitary.

9.8.3 Decomposition
A square matrix A can always be written as a sum of a symmetric A+ and an
antisymmetric matrix A
A = A+ + A (463)
Such a decomposition could e.g. be

A + AT A AT
A= + = A+ + A (464)
2 2

9.9 Toeplitz Matrices

A Toeplitz matrix T is a matrix where the elements of each diagonal is the
same. In the n n square case, it has the following structure:

t11 t12 t1n t0 t1 tn1
.. ..
t21 . . . . . . .. ..

. = t1
. . .
T= . ..
(465)
.. .. .. .. ..
. . t12 . . . t1
tn1 t21 t11 t(n1) t1 t0

A Toeplitz matrix is persymmetric. If a matrix is persymmetric (or orthosym-

metric), it means that the matrix is symmetric about its northeast-southwest
diagonal (anti-diagonal) [12]. Persymmetric matrices is a larger class of matri-
ces, since a persymmetric matrix not necessarily has a Toeplitz structure. There

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 54
9.10 Transition matrices 9 SPECIAL MATRICES

are some special cases of Toeplitz matrices. The symmetric Toeplitz matrix is
given by:
t0 t1 tn1
.. .. ..
t1 . . .
T= .
(466)
.. .. ..
. . t1
tn1 t1 t0
The circular Toeplitz matrix:

t0 t1 tn1
.. .. ..
tn1 . . .
TC =
.
(467)
.. .. ..
. . t1
t1 tn1 t0

The upper triangular Toeplitz matrix:

t0 t1 tn1
..
0 ... ..

. .
TU = . .
, (468)
.. .. ..
. t1
0 0 t0

and the lower triangular Toeplitz matrix:

t0 0 0
. .. .. ..
t1 . .
TL = ..
(469)
.. ..
. . . 0
t(n1) t1 t0

9.9.1 Properties of Toeplitz Matrices

The Toeplitz matrix has some computational advantages. The addition of two
Toeplitz matrices can be done with O(n) flops, multiplication of two Toeplitz
matrices can be done in O(n ln n) flops. Toeplitz equation systems can be solved
in O(n2 ) flops. The inverse of a positive definite Toeplitz matrix can be found
in O(n2 ) flops too. The inverse of a Toeplitz matrix is persymmetric. The
product of two lower triangular Toeplitz matrices is a Toeplitz matrix. More
information on Toeplitz matrices and circulant matrices can be found in [13, 7].

9.10 Transition matrices

A square matrix P is a transition matrix, also known as stochastic matrix or
probability matrix, if
X
0 (P)ij 1, (P)ij = 1
j

The transition matrix usually describes the probability of moving from state i
to j in one step and is closely related to markov processes. Transition matrices

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 55
9.11 Units, Permutation and Shift 9 SPECIAL MATRICES

have the following properties

Prob[i j in 1 step] = (P)ij (470)

Prob[i j in 2 steps] = (P2 )ij (471)
Prob[i j in k steps] = (Pk )ij (472)
n
If all rows are identical P = P (473)
P = , is called invariant (474)
P
where is a so-called stationary probability vector, i.e., 0 i 1 and i i =
1.

9.11 Units, Permutation and Shift

9.11.1 Unit vector
Let ei Rn1 be the ith unit vector, i.e. the vector which is zero in all entries
except the ith at which it is 1.

9.11.2 Rows and Columns

i.th row of A = eTi A (475)

j.th column of A = Aej (476)

9.11.3 Permutations
Let P be some permutation matrix, e.g.

eT2

0 1 0
= eT1
 
P= 1 0 0 = e2 e1 e3 (477)
0 0 1 eT3

For permutation matrices it holds that

PPT = I (478)

and that
eT2 A

PA = eT1 A
 
AP = Ae2 Ae1 Ae3 (479)
eT3 A
That is, the first is a matrix which has columns of A but in permuted sequence
and the second is a matrix which has the rows of A but in the permuted se-
quence.

9.11.4 Translation, Shift or Lag Operators

Let L denote the lag (or translation or shift) operator defined on a 4 4
example by
0 0 0 0
1 0 0 0
L= 0
(480)
1 0 0
0 0 1 0

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 56
9.12 Vandermonde Matrices 9 SPECIAL MATRICES

i.e. a matrix of zeros with one on the sub-diagonal, (L)ij = i,j+1 . With some
signal xt for t = 1, ..., N , the n.th power of the lag operator shifts the indices,
i.e. n 0 for t = 1, .., n
(Ln x)t = (481)
xtn for t = n + 1, ..., N
A related but slightly different matrix is the recurrent shifted operator defined
on a 4x4 example by
0 0 0 1
1 0 0 0
L =
0 1 0 0
(482)
0 0 1 0

i.e. a matrix defined by (L)ij = i,j+1 + i,1 j,dim(L) . On a signal x it has the
effect
(Ln x)t = xt0 , t0 = [(t n) mod N ] + 1 (483)
That is, L is like the shift operator L except that it wraps the signal as if it
was periodic and shifted (substituting the zeros with the rear end of the signal).
Note that L is invertible and orthogonal, i.e.

L1 = LT (484)

9.12 Vandermonde Matrices

A Vandermonde matrix has the form [15]

1 v1 v12 v1n1

1 v2 v22 v2n1
V= . . . (485)

.. ..
.. .. . .
1 vn vn2 vnn1

The transpose of V is also said to a Vandermonde matrix. The determinant is

given by [29] Y
det V = (vi vj ) (486)
i>j

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 57
 10 FUNCTIONS AND OPERATORS

10 Functions and Operators

10.1 Functions and Series
10.1.1 Finite Series
(Xn I)(X I)1 = I + X + X2 + ... + Xn1 (487)

10.1.2 Taylor Expansion of Scalar Function

Consider some scalar function f (x) which takes the vector x as an argument.
This we can Taylor expand around x0
1
f (x)
= f (x0 ) + g(x0 )T (x x0 ) + (x x0 )T H(x0 )(x x0 ) (488)
2
where
f (x)  2 f (x) 
g(x0 ) = H(x0 ) =
x x0 xxT x0
 

10.1.3 Matrix Functions by Infinite Series

As for analytical functions in one dimension, one can define a matrix function
for square matrices X by an infinite series

X
f (X) = cn Xn (489)
n=0

assuming the limit exists and is finite. If the coefficients cn fulfils n cn xn < ,
P
then one can prove that the above series exists and is finite, see [1]. Thus for
any analytical function f (x) there exists a corresponding matrix function f (x)
constructed by the Taylor expansion. Using this one can prove the following
results:
1) A matrix A is a zero of its own characteristic polynomium [1]:
X
p() = det(I A) = cn n p(A) = 0 (490)
n

2) If A is square it holds that [1]

A = UBU1 f (A) = Uf (B)U1 (491)

3) A useful fact when using power series is that

An 0forn if |A| < 1 (492)

10.1.4 Identity and commutations

It holds for an analytical matrix function f (X) that

f (AB)A = Af (BA) (493)

see B.1.2 for a proof.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 58
10.2 Kronecker and Vec Operator 10 FUNCTIONS AND OPERATORS

10.1.5 Exponential Matrix Function

In analogy to the ordinary scalar exponential function, one can define exponen-
tial and logarithmic matrix functions:

X 1 n 1
eA A = I + A + A2 + ... (494)
n=0
n! 2

X 1 1
eA (1)n An = I A + A2 ... (495)
n=0
n! 2

X 1 1
etA (tA)n = I + tA + t2 A2 + ... (496)
n=0
n! 2

X (1)n1 n 1 1
ln(I + A) A = A A2 + A3 ... (497)
n=1
n 2 3

Some of the properties of the exponential function are [1]

eA eB = eA+B if AB = BA (498)
A 1 A
(e ) = e (499)
d tA
e = AetA = etA A, tR (500)
dt
d
Tr(etA ) = Tr(AetA ) (501)
dt
det(eA ) = eTr(A) (502)

10.1.6 Trigonometric Functions

X (1)n A2n+1 1 1
sin(A) = A A3 + A5 ... (503)
n=0
(2n + 1)! 3! 5!

X (1)n A2n 1 1
cos(A) = I A2 + A4 ... (504)
n=0
(2n)! 2! 4!

10.2 Kronecker and Vec Operator

10.2.1 The Kronecker Product
The Kronecker product of an m n matrix A and an r q matrix B, is an
mr nq matrix, A B defined as

A11 B A12 B ... A1n B
A21 B A22 B ... A2n B
AB= (505)

.. ..
. .
Am1 B Am2 B ... Amn B

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 59
10.2 Kronecker and Vec Operator 10 FUNCTIONS AND OPERATORS

The Kronecker product has the following properties (see [19])

A (B + C) = AB+AC (506)
A B 6= BA in general (507)
A (B C) = (A B) C (508)
(A A B B) = A B (A B) (509)
T T T
(A B) = A B (510)
(A B)(C D) = AC BD (511)
(A B)1 = A1 B1 (512)
(A B)+ = A+ B+ (513)
rank(A B) = rank(A)rank(B) (514)
Tr(A B) = Tr(A)Tr(B) = Tr(A B ) (515)
det(A B) = det(A)rank(B) det(B)rank(A) (516)
{eig(A B)} = {eig(B A)} if A, B are square (517)
{eig(A B)} = {eig(A)eig(B)T } (518)
if A, B are symmetric and square
eig(A B) = eig(A) eig(B) (519)

Where {i } denotes the set of values i , that is, the values in no particular
order or structure, and A denotes the diagonal matrix with the eigenvalues of
A.

10.2.2 The Vec Operator

The vec-operator applied on a matrix A stacks the columns into a vector, i.e.
for a 2 2 matrix

  A11
A11 A12 A21
A= vec(A) =
A21 A22 A12
A22

Properties of the vec-operator include (see [19])

vec(AXB) = (BT A)vec(X) (520)

T T
Tr(A B) = vec(A) vec(B) (521)
vec(A + B) = vec(A) + vec(B) (522)
vec(A) = vec(A) (523)
aT XBXT c = vec(X)T (B caT )vec(X) (524)

See B.1.1 for a proof for Eq. 524.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 60
10.3 Vector Norms 10 FUNCTIONS AND OPERATORS

10.3 Vector Norms

10.3.1 Examples

X
||x||1 = |xi | (525)
i
||x||22 = x x H
(526)
" #1/p
X
||x||p = |xi |p (527)
i
||x|| = max |xi | (528)
i

Further reading in e.g. [12, p. 52]

10.4 Matrix Norms

10.4.1 Definitions
A matrix norm is a mapping which fulfils

||A|| 0 (529)
||A|| = 0A=0 (530)
||cA|| = |c|||A||, cR (531)
||A + B|| ||A|| + ||B|| (532)

10.4.2 Induced Norm or Operator Norm

An induced norm is a matrix norm induced by a vector norm by the following

||A|| = sup{||Ax|| | ||x|| = 1} (533)

where || || on the left side is the induced matrix norm, while || || on the right
side denotes the vector norm. For induced norms it holds that
||I|| = 1 (534)
||Ax|| ||A|| ||x||, for all A, x (535)
||AB|| ||A|| ||B||, for all A, B (536)

10.4.3 Examples
X
||A||1 = max |Aij | (537)
j
q i
||A||2 = max eig(AH A) (538)
||A||p = ( max ||Ax||p )1/p (539)
||x||p =1
X
||A|| = max |Aij | (540)
i
sX j
q
||A||F = |Aij |2 = Tr(AAH ) (Frobenius) (541)
ij

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 61
10.5 Rank 10 FUNCTIONS AND OPERATORS

||A||max = max |Aij | (542)

ij
||A||KF = ||sing(A)||1 (Ky Fan) (543)

where sing(A) is the vector of singular values of the matrix A.

10.4.4 Inequalities
E. H. Rasmussen has in yet unpublished material derived and collected the
following inequalities. They are collected in a table as below, assuming A is an
m n, and d = rank(A)
||A||max ||A||1 ||A|| ||A||2 ||A||F ||A||KF
||A||max 1 1 1 1 1
||A||1 m m m m m
||A|| n n n n n
||A||2 mn n m 1 1

||A||F mn n m d 1
||A||KF mnd nd md d d
which are to be read as, e.g.

||A||2 m ||A|| (544)

10.4.5 Condition Number

p
The 2-norm of A equals (max(eig(AT A))) [12, p.57]. For a symmetric, pos-
itive definite matrix, this reduces to max(eig(A)) The condition number based
on the 2-norm thus reduces to
max(eig(A))
kAk2 kA1 k2 = max(eig(A)) max(eig(A1 )) = . (545)
min(eig(A))

10.5 Rank
10.5.1 Sylvesters Inequality
If A is m n and B is n r, then

rank(A) + rank(B) n rank(AB) min{rank(A), rank(B)} (546)

10.6 Integral Involving Dirac Delta Functions

Assuming A to be square, then
Z
1
p(s)(x As)ds = p(A1 x) (547)
det(A)

Assuming A to be underdetermined, i.e. tall, then

( )
Z 1 T p(A+ x) if x = AA+ x
p(s)(x As)ds = det(A A) (548)
0 elsewhere

See [9].

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 62
10.7 Miscellaneous 10 FUNCTIONS AND OPERATORS

10.7 Miscellaneous
For any A it holds that

rank(A) = rank(AT ) = rank(AAT ) = rank(AT A) (549)

It holds that

A is positive definite B invertible, such that A = BBT (550)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 63
 A ONE-DIMENSIONAL RESULTS

A One-dimensional Results
A.1 Gaussian
A.1.1 Density
(x )2
 
1
p(x) = exp (551)
2 2 2 2

A.1.2 Normalization
Z
(s)2
e 22 ds = 2 2 (552)
r  2 
b 4ac
Z
(ax2 +bx+c)
e dx = exp (553)
a 4a
r  2 
4c2 c0
Z
2 c
ec2 x +c1 x+c0 dx = exp 1 (554)
c2 4c2

A.1.3 Derivatives
p(x) (x )
= p(x) (555)
2
ln p(x) (x )
= (556)
2
1 (x )2
 
p(x)
= p(x) 1 (557)
2
1 (x )2
 
ln p(x)
= 1 (558)
2

A.1.4 Completing the Squares

c2 x2 + c1 x + c0 = a(x b)2 + w
1 c1 1 c21
a = c2 b= w= + c0
2 c2 4 c2
or
1
c2 x2 + c1 x + c0 = (x )2 + d
2 2
c1 1 c21
= 2 = d = c0
2c2 2c2 4c2

A.1.5 Moments
If the density is expressed by

(s )2
 
1
p(x) = exp or p(x) = C exp(c2 x2 + c1 x) (559)
2 2 2 2

then the first few basic moments are

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 64
A.2 One Dimensional Mixture of Gaussians
A ONE-DIMENSIONAL RESULTS

c1
hxi = = 2c2  2
1 c1
hx2 i = 2 + 2 = 2c2 + h 2c2
c21
i
c1
hx3 i = 3 2 + 3 = (2c ) 2 3 2c2
 2 4  2    2
c1 c1 1 1
hx4 i = 4 + 62 2 + 3 4 = 2c2 + 6 2c2 2c2 +3 2c2

and the central moments are

h(x )i = 0 = h0 i
2 2 1
h(x ) i = = 2c2
h(x )3 i = 0 = 0
h i2
1
h(x )4 i = 3 4 = 3 2c2

A kind of pseudo-moments (un-normalized integrals) can easily be derived as

Z r  2 
c1
exp(c2 x2 + c1 x)xn dx = Zhxn i = exp hxn i (560)
c2 4c2

From the un-centralized moments one can derive other entities like
1
hx2 i hxi2 = 2 = 2c2
2c1
hx3 i hx2 ihxi = 2 2 = (2c2 )2
c2
h i
2
hx4 i hx2 i2 = 2 4 + 42 2 = (2c2 )2 1 4 2c12

A.2 One Dimensional Mixture of Gaussians

A.2.1 Density and Normalization
K
1 (s k )2
 
X k
p(s) = exp (561)
k2
p
k
2k2 2

A.2.2 Moments
A useful fact of MoG, is that
X
hxn i = k hxn ik (562)
k

where hik denotes average with respect to the k.th component. We can calculate
the first four moments from the densities
1 (x k )2
 
X 1
p(x) = k p exp (563)
2k2 2 k2
k
X
k Ck exp ck2 x2 + ck1 x
 
p(x) = (564)
k

as

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 65
 B PROOFS AND DETAILS

h i
P P ck1
hxi = k k k = k k
 2ck2  2 
2 2 1 ck1
2k )
P P
hx i = k k (k + = k k 2ck2 + 2ck2
c2k1
h h ii
ck1
hx3 i 2 3
P P
= k k (3k k + k ) = k k (2ck2 )2 3 2ck2
 2  2 
1 ck1 c2k1
hx4 i 4 2 2 4
P P
= k k (k + 6k k + 3k ) =
k k 2ck2 2ck2 6 2ck2 + 3

If all the gaussians are centered, i.e. k = 0 for all k, then

hxi = 0 = 0 h i
1
hx2 i 2
P P
= k k k = k k 2ck2
hx3 i = 0 = 0
h i2
1
hx4 i k 3k4
P P
= k = k k 3 2ck2

From the un-centralized moments one can derive other entities like
 2 
hx2 i hxi2 2
P
= k,k0 k k k + k k k
0 0

hx3 i hx2 ihxi = 2 3 2 2
P
k,k 0 k k 0 3k k + k (k + k )k 0
 4 
hx4 i hx2 i2 2 2 4 2 2 2 2
P
= k,k0 k k k + 6k k + 3k (k + k )(k0 + k0 )
0

A.2.3 Derivatives
Defining p(s) = k k Ns (k , k2 ) we get for a parameter j of the j.th compo-
P
nent
ln p(s) j Ns (j , j2 ) ln(j Ns (j , j2 ))
=P 2 (565)
j k k Ns (k , k ) j
that is,
ln p(s) j Ns (j , j2 ) 1
= P 2 (566)
j k k Ns (k , k ) j
ln p(s) j Ns (j , j2 ) (s j )
= 2 (567)
j2
P
j k k Ns (k , k )
" #
ln p(s) j Ns (j , j2 ) 1 (s j )2
= 2 1 (568)
j2
P
j k k Ns (k , k ) j

Note thatP k must be constrained to be proper ratios. Defining the ratios by

j = erj / k erk , we obtain
ln p(s) X ln p(s) l l
= where = l (lj j ) (569)
rj l rj rj
l

B Proofs and Details

B.1 Misc Proofs
B.1.1 Proof of Equation 524
The following proof is work of Florian Roemer. Note the the vectors and ma-
trices below can be complex and the notation XH is used for transpose and
conjugated, while XT is only transpose of the complex matrix.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 66
B.1 Misc Proofs B PROOFS AND DETAILS

Define the row vector y = aH XB and the column vector z = XH c. Then

aT XBXT c = yz = zT yT

Note that y can be rewritten as vec(y)T which is the same as

vec(conj(y))H = vec(aT conj(X)conj(B))H

where conj means complex conjugated. Applying the vec rule for linear forms
Eq 520, we get

y = (BH aT vec(conj(X))H = vec(X)T (B conj(a))

where we have also used the rule for transpose of Kronecker products. For yT
this yields (BT aH )vec(X). Similarly we can rewrite z which is the same as
vec(zT ) = vec(cT conj(X)). Applying again Eq 520, we get

z = (I cT )vec(conj(X))

where I is the identity matrix. For zT we obtain vec(X)(I c). Finally, the
original expression is zT yT which now takes the form

vec(X)H (I c)(BT aH )vec(X)

the final step is to apply the rule for products of Kronecker products and by
that combine the Kronecker products. This gives

vec(X)H (BT caH )vec(X)

which is the desired result.

B.1.2 Proof of Equation 493

For any analytical function f (X) of a matrix argument X, it holds that

!
X
n
f (AB)A = cn (AB) A
n=0

X
= cn (AB)n A
n=0
X
= cn A(BA)n
n=0
X
= A cn (BA)n
n=0
= Af (BA)

B.1.3 Proof of Equation 91

Essentially we need to calculate

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 67
B.1 Misc Proofs B PROOFS AND DETAILS

(Xn )kl X
= Xk,u1 Xu1 ,u2 ...Xun1 ,l
Xij Xij u1 ,...,un1
= k,i u1 ,j Xu1 ,u2 ...Xun1 ,l
+Xk,u1 u1 ,i u2 ,j ...Xun1 ,l
..
.
+Xk,u1 Xu1 ,u2 ...un1 ,i l,j
n1
X
= (Xr )ki (Xn1r )jl
r=0
n1
X
= (Xr Jij Xn1r )kl
r=0

Using the properties of the single entry matrix found in Sec. 9.7.4, the result
follows easily.

B.1.4 Details on Eq. 571

det(XH AX) = det(XH AX)Tr[(XH AX)1 (XH AX)]

= det(XH AX)Tr[(XH AX)1 ((XH )AX + XH (AX))]
= det(XH AX) Tr[(XH AX)1 (XH )AX]
+Tr[(XH AX)1 XH (AX)]

= det(XH AX) Tr[AX(XH AX)1 (XH )]

+Tr[(XH AX)1 XH A(X)]

First, the derivative is found with respect to the real part of X

det(XH AX)  Tr[AX(XH AX)1 (XH )]
= det(XH AX)
<X <X
Tr[(XH AX)1 XH A(X)] 
+
<X
= det(XH AX) AX(XH AX)1 + ((XH AX)1 XH A)T

Through the calculations, (100) and (240) were used. In addition, by use of
(241), the derivative is found with respect to the imaginary part of X
det(XH AX)  Tr[AX(XH AX)1 (XH )]
i = i det(XH AX)
=X =X
Tr[(XH AX)1 XH A(X)] 
+
=X
= det(X AX) AX(XH AX)1 ((XH AX)1 XH A)T
H

Hence, derivative yields

det(XH AX) 1  det(XH AX) det(XH AX) 
= i
X 2 <X =X
H H 1 H

T
= det(X AX) (X AX) X A

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 68
B.1 Misc Proofs B PROOFS AND DETAILS

and the complex conjugate derivative yields

det(XH AX) 1  det(XH AX) det(XH AX) 

= +i
X 2 <X =X
H H 1
= det(X AX)AX(X AX)

Notice, for real X, A, the sum of (249) and (250) is reduced to (54).
Similar calculations yield

det(XAXH ) 1  det(XAXH ) det(XAXH ) 

= i
X 2 <X =X
H H H 1 T


= det(XAX ) AX (XAX ) (570)

and

det(XAXH ) 1  det(XAXH ) det(XAXH ) 

= +i
X 2 <X =X
= det(XAXH )(XAXH )1 XA (571)

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 69
REFERENCES REFERENCES

References
[1] Karl Gustav Andersson and Lars-Christer Boiers. Ordinaera differentialek-
vationer. Studenterlitteratur, 1992.
[2] Jorn Anemuller, Terrence J. Sejnowski, and Scott Makeig. Complex inde-
pendent component analysis of frequency-domain electroencephalographic
data. Neural Networks, 16(9):13111323, November 2003.
[3] S. Barnet. Matrices. Methods and Applications. Oxford Applied Mathe-
matics and Computin Science Series. Clarendon Press, 1990.
[4] Christopher Bishop. Neural Networks for Pattern Recognition. Oxford
University Press, 1995.
[5] Robert J. Boik. Lecture notes: Statistics 550. Online, April 22 2002. Notes.
[6] D. H. Brandwood. A complex gradient operator and its application in
adaptive array theory. IEE Proceedings, 130(1):1116, February 1983. PTS.
F and H.
[7] M. Brookes. Matrix Reference Manual, 2004. Website May 20, 2004.
[8] Contradsen K., En introduktion til statistik, IMM lecture notes, 1984.
[9] Mads Dyrholm. Some matrix results, 2004. Website August 23, 2004.
[10] Nielsen F. A., Formula, Neuro Research Unit and Technical university of
Denmark, 2002.
[11] Gelman A. B., J. S. Carlin, H. S. Stern, D. B. Rubin, Bayesian Data
Analysis, Chapman and Hall / CRC, 1995.
[12] Gene H. Golub and Charles F. van Loan. Matrix Computations. The Johns
Hopkins University Press, Baltimore, 3rd edition, 1996.
[13] Robert M. Gray. Toeplitz and circulant matrices: A review. Technical
report, Information Systems Laboratory, Department of Electrical Engi-
neering,Stanford University, Stanford, California 94305, August 2002.
[14] Simon Haykin. Adaptive Filter Theory. Prentice Hall, Upper Saddle River,
NJ, 4th edition, 2002.
[15] Roger A. Horn and Charles R. Johnson. Matrix Analysis. Cambridge
University Press, 1985.
[16] Mardia K. V., J.T. Kent and J.M. Bibby, Multivariate Analysis, Academic
Press Ltd., 1979.
[17] Mathpages on Eigenvalue Problems and Matrix Invariants,

http://www.mathpages.com/home/kmath128.htm

[18] Carl D. Meyer. Generalized inversion of modified matrices. SIAM Journal

of Applied Mathematics, 24(3):315323, May 1973.

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 70
REFERENCES REFERENCES

[19] Thomas P. Minka. Old and new matrix algebra useful for statistics, De-
cember 2000. Notes.

[20] Daniele Mortari OrthoSkew and OrthoSym Matrix Trigonometry John

Lee Junkins Astrodynamics Symposium, AAS 03265, May 2003. Texas
A&M University, College Station, TX
[21] L. Parra and C. Spence. Convolutive blind separation of non-stationary
sources. In IEEE Transactions Speech and Audio Processing, pages 320
327, May 2000.
[22] Kaare Brandt Petersen, Jiucang Hao, and Te-Won Lee. Generative and
filtering approaches for overcomplete representations. Neural Information
Processing - Letters and Reviews, vol. 8(1), 2005.
[23] John G. Proakis and Dimitris G. Manolakis. Digital Signal Processing.
Prentice-Hall, 1996.
[24] Laurent Schwartz. Cours dAnalyse, volume II. Hermann, Paris, 1967. As
referenced in [14].
[25] Shayle R. Searle. Matrix Algebra Useful for Statistics. John Wiley and
Sons, 1982.
[26] G. Seber and A. Lee. Linear Regression Analysis. John Wiley and Sons,
2002.
[27] S. M. Selby. Standard Mathematical Tables. CRC Press, 1974.

[28] Inna Stainvas. Matrix algebra in differential calculus. Neural Computing

Research Group, Information Engeneering, Aston University, UK, August
2002. Notes.
[29] P. P. Vaidyanathan. Multirate Systems and Filter Banks. Prentice Hall,
1993.

[30] Max Welling. The Kalman Filter. Lecture Note.

[31] Wikipedia on minors: Minor (linear algebra),

http://en.wikipedia.org/wiki/Minor_(linear_algebra)

[32] Zhaoshui He, Shengli Xie, et al, Convolutive blind source separation in
frequency domain based on sparse representation, IEEE Transactions on
Audio, Speech and Language Processing, vol.15(5):1551-1563, July 2007.
[33] Karim T. Abou-Moustafa On Derivatives of Eigenvalues and Eigenvectors
of the Generalized Eigenvalue Problem. McGill Technical Report, October
2010.
[34] Mohammad Emtiyaz Khan Updating Inverse of a Matrix When a Column
is Added/Removed. Emt CS,UBC February 27, 2008

Petersen & Pedersen, The Matrix Cookbook, Version: November 15, 2012, Page 71
Index
Anti-symmetric, 54 Normal-Inverse Gamma distribution, 37
Normal-Inverse Wishart distribution, 39
Block matrix, 46
Orthogonal, 49
Chain rule, 15
Cholesky-decomposition, 32 Power series of matrices, 58
Co-kurtosis, 34 Probability matrix, 55
Co-skewness, 34 Pseudo-inverse, 21
Condition number, 62
Cramers Rule, 29 Schur complement, 41, 47
Single entry matrix, 52
Derivative of a complex matrix, 24 Singular Valued Decomposition (SVD),
Derivative of a determinant, 8 31
Derivative of a trace, 12 Skew-Hermitian, 48
Derivative of an inverse, 9 Skew-symmetric, 54
Derivative of symmetric matrix, 15 Stochastic matrix, 55
Derivatives of Toeplitz matrix, 16 Student-t, 37
Dirichlet distribution, 37 Sylvesters Inequality, 62
Symmetric, 54
Eigenvalues, 30
Eigenvectors, 30 Taylor expansion, 58
Exponential Matrix Function, 59 Toeplitz matrix, 54
Transition matrix, 55
Gaussian, conditional, 40 Trigonometric functions, 59
Gaussian, entropy, 44
Gaussian, linear combination, 41 Unipotent, 49
Gaussian, marginal, 40
Gaussian, product of densities, 42 Vandermonde matrix, 57
Generalized inverse, 21 Vec operator, 59, 60

Hadamard inequality, 52 Wishart distribution, 38

Hermitian, 48 Woodbury identity, 18

Idempotent, 49

Kronecker product, 59

LDL decomposition, 33
LDM-decomposition, 33
Linear regression, 28
LU decomposition, 32
Lyapunov Equation, 30

Moore-Penrose inverse, 21
Multinomial distribution, 37

Nilpotent, 49
Norm of a matrix, 61
Norm of a vector, 61

72