6-ConceptsProbability Page 195 Wednesday, February 4, 2004 3:00 PM
Concepts of Probability 195
EXHIBIT 6.1 Graph of a Normal Variable with Zero Mean and σ = 100
Multivariate normal distributions are characterized by the same
exponential functional form. However, a multivariate normal distribu-
tion in nvariables is identified by nmeans, one for each axis, and by a
n×nsymmetrical variance-covariance matrix. For instance, a bivariate
normal distribution is characterized by two expected values, two vari-
ances and one covariance. We can write the general expression of a
bivariate normal distribution as follows:
1
exp (^) – ---Q
2
fxy( , ) = -----------------------------------------
2 πσXσY 1 – ρ
2
1 x– μX
(^2) x– μ
X y– μY y– μY
(^2)
Q= --------------- ---------------- – 2 ρ ---------------- --------------- + ---------------
1 – ρ^2 σX σX σY σY
where ρ is the correlation coefficient.