Professional Documents
Culture Documents
y Bz d (6.3)
where B AC and d Aμ b .
Equation (6.3) is a linear transform of standard Normal random variables.
Definition 6.1
A random vector X ( X1 , X 2 , , X k ) is said to have a Multivariate Normal Distribution if x
is a linear transform of a vector Z (Z1 , Z 2 , , Z m ) of independent standard Normal
variables
2 THE MULTIVARIATE NORMAL DISTRIBUTION
X AZ μ (6.4)
Exercise
Find E(X) and D(X) in the following linear transformation of the standard normal vector
Z (Z1 , Z 2 , Z 3 )
X 1 Z1 2Z 2 Z 3 2
X 2 3Z1 2Z 2
Remarks 6.1
Notation
A multivariate Normal distribution with mean vector μ and variance-covariance matrix Σ will
be designated as N (μ, Σ) k , where k is the dimension of the distribution.
3
1. Linear transform
If Y AX b , where A is a p k matrix of constants, b is a constant vector, and X
~ N (μ, Σ) k then
(a) Y ~ N (Aμ b, AΣA)
(b) If the distribution of X is non-singular and Y (Y1 , Y2 , , Yp ) , then the
distribution of Y is also non-singular if and only if rank(A) p
Proof
Denote Σ ( ij )
If X i , i 1, 2,, k are mutually independent, then they are pair-wisely independent,
and pair-wisely uncorrelated. This implies that ij 0 , i j . Therefore, Σ is diagonal
Now suppose that Σ is diagonal..We note that
X AZ μ , where N 0, Ik and Σ AA .
We use the property of multivariate moment generating functions (mmgf) to establish
the independence.
Now assume that Σ is diagonal and given as Σ diag( 11 , 22 , , kk )
Let X ( X 1 , X 2 , , X k ) and Z (Z1 , Z 2 , , Z m )
The mmgf of Z is given by
M z (t) exp12 tt where t (t1 , t 2 , , t m ) and define s (s1 , s2 , , sk ) . The mmgf
of X is given by
e μs M AZ (s)
e μs M Z (As)
Making substitutions and simplifying gives
expμs 12 sΣ X s
k k
M X (s) exp j s j 12 jj s 2j
j 1 j 1
exp j s j 12 jj s 2j
k
j 1
Now
M X j (s) exp j s j 12 jj s 2j implying that X j ~ N ( j , jj ) . Therefore,
k
M X (s) M X j (s j )
j 1
Therefore, since the mmgf of X is the product of the mgfs of the components X j , it
means that X j j 1, 2, , k are independent.
Proof
Λ diag( 1 , 2 , , k ) diag( 1 , 2 , , k )
Λ 2Λ
1 1
2
Thus,
5
Σ T Λ T
SS
non-singular.
D(Z) S 1 D(X μ) (S 1 )
S 1 D(X) (S 1 )
S 1 Σ (S 1 )
S 1 (SS) (S 1 )
S 1S S (S) 1
Ik
Example 6.1
Let X ( X1 , X 2 , X 3 ) have multivariate Normal distribution with mean vector (1, 0, 2)
and variance-covariance matrix
2 0 2
Σ 0 1 1
2 1 5
Solution
4 2 2
We find that Σ 4 and Σ 1 2 6 2 .
1
4
2 2 2
Q(x) ( x1 1) 2 32 ( x2 0) 2 12 ( x3 2) 2 ( x1 1) x2 ( x1 1)(x3 2) x2 ( x3 2)
This simplifies as
In Example (6.1), we are able to obtain the joint density function of ( X 1 , X 2 , X 3 ) (or the
quadratic form, Q(x) ) since the variance-covariance matrix and the mean vector were provided.
Conversely, it should be possible to derive the mean vector and variance-covariance matrix given
the quadratic form.
c 2Aμ (6.11)
1
μ Σc (6.12)
2
For example, given the quadratic form in the solution to Example 6.1, obtain the mean vector
and variance-covariance matrix.
8 THE MULTIVARIATE NORMAL DISTRIBUTION
Review Exercise 6
1. Find the mean vector and variance-covariance matrix of the multivariate Normal distribution
with density (x) K exp 12 Q(x), if
5 0 5
Σ 0 5 5
1
2
5 5 10
Find the density function of X.
y1 x1 2 x2 x3
y 2 3x1 2 x2 2 x3
y3 2 x1 x2 4 x3
Find
(a) the multivariate moment generating function of Y;
(b) the probability density function for Y.
(c) Comment on your results in (a) and (b).
REFERENCES