Probability Theory and Random
Processes (MA225)
Lecture Slides
Lecture 19
Bivariate normal
Def: A two dimensional random vector X = (X , Y ) is said to have a
bivariate normal distribution if aX + bY is a univariate normal for all
(a, b) ∈ R2 \ (0, 0).
Theorem: If µ = E (X ) and Σ is the variance-covariance matrix of X ,
then for any fixed u = (a, b) ∈ R2 \ (0, 0), u ′ X ∼ N(u ′ µ, u ′ Σu).
Theorem: Let X be a bivariate normal random vector, then
′ 1 ′
MX (t) = e t µ+ 2 t Σt for all t ∈ R2 .
Remark: Thus the bivariate normal distribution is completely specified
by the mean vector µ and the variance-covariance matrix Σ. We may
therefore denote a bivariate normal distribution by N2 (µ, Σ).
Def: A two dimensional random vector X is said to have a bivariate
normal distribution if it can be expressed in the form X = µ + AY ,
where A is a 2 × 2 matrix of real numbers, Y = (Y1 , Y2 ) and Y1 and
Y2 are i.i.d N(0, 1). In this case E (X ) = µ and Σ = AA′ .
Theorem: If X ∼ N2 (µ, Σ), then X ∼ N(µ1 , σ11 ) and
Y ∼ N(µ2 , σ22 ).
Remark: The converse of the above theorem is not true.
Theorem: Let X ∼ N2 (µ, Σ) be such that Σ is invertible, then, for
all x ∈ R2 , X has a joint PDF given by
1 1 ′ −1
f (x) = exp − (x − µ) Σ (x − µ)
2π|Σ|1/2 2
1
= p e A(x, y , µx , µy , σx , σy , ρ) ,
2πσx σy 1 − ρ2
where
( 2 2 )
1 x − µx x − µx y − µy y − µy
A=− − 2ρ + .
2(1 − ρ2 ) σx σx σy σy
Remark: If Cov (X , Y ) = 0, then X and Y are independent.
Theorem: Let X ∼ N2 (µ, Σ) be such that Σ is invertible, then
1 for all y ∈ R, the conditional PDF of X given Y = y is given by
" 2 #
1 x − µx|y
1
fX |Y (x|y ) = √ exp − for x ∈ R,
σx|y 2π 2 σx|y
where µx|y = µx + ρ σσyx (y − µy ) and σx|y
2
= σx2 (1 − ρ2 ).
2 E (X |Y = y ) = µx|y = µx + ρ σσyx (y − µy ) for all y ∈ R.
Theorem:
Pn Let X1 , X2 , . . . Xn be i.i.d. N(0, 1) random variables. Then
2 2
i=1 Xi ∼ Gamma(n/2, 1/2) ≡ χn .
Theorem: Let X1 , X2 , . . . Xn be i.i.d. N(µ, σ 2 ) random variables.
2
Then X ∼ N(µ, σ 2 /n), (n−1)Sσ2
∼ χ2n−1 , and X and S 2 are
independently distributed. Here X = n1 ni=1 Xi and
P
1
Pn
S 2 = n−1 2
i=1 (Xi − X ) .
