Geoscience Reference
In-Depth Information
A.7.2.6 The Lognormal Distribution
lnX has a normal distribution.
The density of the distribution of X is then
X has a lognormal distribution
↔
x
p
Þ
1
2
f
1
2
f
X
x
ðÞ
¼
ð
2
p
exp½
ð
ln
ð
1
=
2
x
=nÞÞ
=f
1R
þ
x
ðÞ;
for
ʾ
the median of the distribution of X and
ʶ
= Var(lnX).
2
EX ¼
n
exp½ 1
=ð f
;
2
2
Var X
ðÞ
¼
ðÞ
EX
½exp
ðf
Þ
1
:
A.7.3 Examples of Multidimensional Distributions
For joint distributions, specially useful is to consider joint normal distributions.
A.7.3.1 Bidimensional Normal Distribution
The pair of random variables (X, Y) has a bidimensional normal distribution with
expected values
2
X
2
Y
ʼ
X
and
ʼ
Y
, variances
r
and
r
and correlation coef
cient
ˁ
XY
,
what is formally denoted by (X,Y)
Normal(
l
X
; l
Y
; r
2
X
; r
2
Y
; q
XY
) if and only if
*
their joint density has the form
2
2
f
XY
x
ðÞ
¼
;
y
C
1
exp
f
C
2
½½
ð
x
l
X
Þ=r
X
2
q
½
ð
x
l
X
Þ=r
X
½
ð
y
l
Y
Þ=r
Y
þ
½
ð
y
l
Y
Þ=r
Y
g
where
pr
X
r
Y
Þ
1
XY
Þ
1
=
2
2
C
1
¼
ð
2
ð
1
q
and
2
XY
1
C
2
¼½
2
ð
1
q
:
So,
if
q
¼
0
;
X and Y are independent
:
This extends to vectors of random variables.
Search WWH ::
Custom Search