Digital Signal Processing Reference
In-Depth Information
is a discrete random variable with only two discrete values 1 and 0, that has the
corresponding probabilities,
PfX ¼
1
g¼p:
(5.102)
PfX ¼
0
g¼
1
p ¼ q:
(5.103)
5.6.2 What Is a Binomial Random Variable?
Consider
n
independent random Bernoulli exp
e
riments. As a result of each experi-
ment, either event
A
or its complement
A
occurs, with the corresponding
probabili
tie
s (
5.100
) and (
5.101
). The out
co
mes of repeated trials are sequences
of
A
and
A
where order of sequences
A
or
A
is arbitrary.
AAAA
...
AAAA
...
A A:
(5.104)
The number of
A
events in
n
repeated Bernoulli trials can be any integer, from
0to
n
. A random variable associated with the number of the occurrences of event
A
is said to be a
binomial random variable
, with the parameters,
n
and
p
, where
p
is a
probability of event
A
.
A binomial random variable
X
is a discrete random variable with the discrete
values
x ¼
0,
...
,
n
. In order to describe this variable, we need the probability
PfX ¼ kg¼P
X
ðk
;
nÞ;
k ¼
0
;
...
; n:
(5.105)
The total number of the combinations of
k
events
A
and (
n k
) events
A
in
n
repeated experiments is equal to
¼
n
k
n
ð
n
1
Þ;
...
; ð
n
k
þ
1
Þ
1
2
3
k
n
!
k
!
ðn kÞ
!
C
n
¼
¼
;
(5.106)
where
C
n
is a binomial coefficient and the symbol “
N
!” means factorial:
N
!
¼
1
2
3
ðN
1
ÞN:
(5.107)
Keeping in mind that the
e
xperiments are independent, each combination of
k
events
A
and (
n k
) events
A
, will have a probability,
p
k
q
nk
;
(5.108)
resulting in:
p
k
q
nk
n
k
P
X
ðk
;
nÞ¼C
n
p
k
q
nk
¼
:
(5.109)
This expression is called a
Bernoulli formula
and
Binomial probability law
[LEO94, p. 62].
Search WWH ::
Custom Search