Multivariate Normal
Multivariate Normal
distribution
The bivariate normal distribution
where
Q(x, y ) = (x − µ)T Σ−1 (x − µ)
and
x µx
x= , µ= .
y µy
If random variables (X , Y ) have joint probability density given by
fX ,Y above, then we say that (X , Y ) have a bivariate normal
distribution and write
(X , Y )T ∼ N2 (µ, Σ).
x T Ax ≥ 0
σx = σy, ρ = 0
3
90%
2
y
−1
80%
−2
95%
−3 99%
−3 −2 −1 0 1 2 3
x
σx = 2σy, ρ = 0
3
0
y
Pictures −1
80%
−2
95%
−3 99%
−3 −2 −1 0 1 2 3
x
σx = 2σy, ρ = 0
3
2
99%
1 80%
y
−1 90%
95%
−2
−3
−3 −2 −1 0 1 2 3
x
2σx = σy, ρ = 0
3 99%
−1 90%
95%
Pictures −2
−3
−3 −2 −1 0 1 2 3
x
2σx = σy, ρ = 0
3 99%
95%
2
80%
y
−1
−2 90%
−3
−3 −2 −1 0 1 2 3
x
Pictures
σx = σy, ρ = 0.75
3
90%
2
y
−1
80%
−2
95%
−3 99%
−3 −2 −1 0 1 2 3
x
σx = σy, ρ = − 0.75
3 99%
0
y
Pictures −1
80%
−2
95%
−3 99%
−3 −2 −1 0 1 2 3
x
σx = σy, ρ = − 0.75
3 99%
90%
2
y
−1
80%
−2
95%
−3
−3 −2 −1 0 1 2 3
x
95%
−3
−3 −2 −1 0 1 2 3
x
y
−1
90%
−2
99%
−3
−3 −2 −1 0 1 2 3
x
Comments
Solution
The marginal distribution of X is X ∼ N(2, 12 ).
Using the formula for the conditional
σy
Y |X = x ∼ N(µy + ρ (x − µx ), σy2 (1 − ρ2 ))
σx
∼ N(3 + 0.5(x − 2), 0.75).
Example 5.1
Solution
The marginal distribution of X is X ∼ N(2, 12 ).
Using the formula for the conditional
σy
Y |X = x ∼ N(µy + ρ (x − µx ), σy2 (1 − ρ2 ))
σx
∼ N(3 + 0.5(x − 2), 0.75).
Simulation results
1 n p t s = 500
2 x = rnorm ( n p t s , mean=2, s d = 1 )
3 y = rnorm ( n p t s , mean=3+0.5∗ ( x −2) , s d=s q r t ( 0 . 7 5 ) )
●
6
● ●
● ●
● ● ●●
5
● ● ● ●
● ●●●
● ● ● ●● ● ●
●
● ● ●● ● ● ● ●
●
●●●● ● ● ● ● ●
●● ● ● ●●
● ● ●●●● ● ● ●
4
●●●●● ● ●
●●●●● ●●●
● ●●●● ●
●● ●●●● ●
● ●● ●● ● ●
● ●● ● ●●●● ● ● ●●●● ●●●●●● ● ● ●
● ● ●● ● ●
● ● ● ●
● ●● ●● ●●
● ●●●●●●●● ●●●●●●●●●●●
●●●● ●
●● ●●● ●●●● ●
● ●●●●
●
● ●
●
●● ●●● ● ● ●●
● ●
●● ●
●● ●●●●●●● ●●●
● ● ●● ●●● ● ● ●
● ●● ● ●● ● ● ●● ● ●● ● ●● ●
y
● ●●●
3
● ● ●
● ● ● ● ●
● ●
●●● ● ●●●●●● ●● ● ●● ●●● ●
● ●
● ●●● ●● ●●● ●● ●●
● ●● ● ●● ● ● ●● ● ●●
● ●●●●● ●●
●●●●● ●●
●●
● ● ●●●●●●● ●● ● ● ●
●●
● ●● ● ●●● ● ● ●●●● ●●●
●●● ● ● ●●●● ●
●●●● ● ● ●●● ●●
●
● ● ●●● ● ●●
● ● ● ●● ● ●● ● ●●●
● ● ● ● ●● ●
2
●●
● ●● ●● ●● ●●●● ●●●● ● ● ●
● ●● ● ● ●
●
● ● ● ●● ● ●● ●●● ● ●
● ●●●●
● ●
● ● ● ●
● ● ●● ●
●● ● ● ●
● ● ● ● ● ● ●●
1
●
●●
● ● ●
●●●
●
●
● ●
0
0 1 2 3 4 5
x
Probability calculation
Suppose
X 4 8 2
∼ N2 , .
Y 1 2 5
The random variable Z is defined by Z = X + 3Y . What is the
distribution of Z ?
Extra example
E [Z ] = 1 × µx + 3 × µy = 1 × 4 + 3 × 1 = 7.
E [Z ] = 1 × µx + 3 × µy = 1 × 4 + 3 × 1 = 7.
E [Z ] = 1 × µx + 3 × µy = 1 × 4 + 3 × 1 = 7.
E [Z ] = 1 × µx + 3 × µy = 1 × 4 + 3 × 1 = 7.
−20 0 20 40
The multivariate normal distribution
where
Q(x) = (x − µ)T Σ−1 (x − µ).
This definition makes sense for any column vector µ ∈ Rn and any
positive definite n × n matrix Σ.
Remarks