Digital Signal Processing Reference
In-Depth Information
5.6.3 Binomial Distribution and Density
From (
5.108
) the
binomial distribution
is given as:
p
k
q
nk
uðx kÞ¼
X
F
X
ðxÞ¼
X
n
n
n
k
P
X
ðk
;
nÞuðx kÞ:
(5.110)
k¼
0
k¼
0
The probabilities
P
X
(
k
;
n
) are called the
probability mass function
, as indicated
Keeping in mind that a binomial variable is a discrete variable, its corresponding
density function is:
p
k
q
nk
dðx kÞ¼
X
f
X
ðxÞ¼
X
n
n
n
k
P
X
ðk; nÞdðx kÞ:
(5.111)
k¼
0
k¼
0
The name “binomial” comes up from the probability (
5.109
) which itself
represents the
n
th degree of the Newton binomial:
¼
X
n
n
C
n
p
k
q
nk
ðp þ qÞ
:
(5.112)
k¼
0
Taking into account (
5.101
), and using (
5.109
) from (
5.112
), we arrive at:
¼
1
¼
X
¼
X
n
n
1
n
C
n
p
k
q
nk
P
X
ðk; nÞ:
(5.113)
k¼
0
k¼
0
The expression (
5.113
) proves that the binomial PDF satisfies the condition in
(
2.83
).
Additionally, the expression (
5.112
) can be useful in calculating the characteris-
tic function, as shown below.
5.6.4 Characteristic Functions and Moments
The characteristic function of a binomial variable is:
f
X
ðoÞ¼Ef
e
joX
g
¼
X
e
jok
P
X
ðk
;
nÞ¼
X
¼
X
n
n
n
k
e
jok
C
n
p
k
q
nk
C
n
ðp
e
jo
q
nk
Þ
:
(5.114)
k¼
0
k¼
0
k¼
0
Search WWH ::
Custom Search