D ist r ibu si Var iab el A cak D isk r i t
5.4 D ist r ibu si Var iab el A cak D isk r i t
Here we are going t o st udy a few discret e random variable dist ribut ions.
5.4.1 B er noul l i t r i al s dan di st r i busi B ernoul l i
² A Bernoulli t rail is an experiment wit h result of success or failure. ² We can use a random var iable t o model t his phenomena. Let X j =1
if it is a success, X j = 0 if it is a failure. ² A consecut ive n Bernoulli t rails are called a Bernoulli process if
– t he t r ails are independent of each ot her; – each t rail has only t wo possible out comes (success or failure, t rue
² The following relat ions hold. p(x 1 ;x 2 ;:::;x n )=p 1 (x 1 ) ¢p 2 (x 2 ) ¢: : : p n (x n )
which means t he probabilit y of t he result of a sequence of event s is equal t o t he product of t he probabilit ies of each event .
Because t he event s are independent and t he probabilit y remains t he same,
x j = 1; j = 1; 2; : : : ; n p j (x j ) = p(x j )= 1¡p=qx j = 0; j = 1; 2; : : : ; n
² Not e t hat t he ” locat ion” of t he p i ’s don’t mat t er. I t is t he count of p i ’s t hat is import ant .
² Examples of Bernoulli t rails include: a conscut ive t hrowing of a ” fair” coin, count ing heads and t ails; a pass or fail t est on a sequence of a part icular component s of t he ” same” qualit y; and ot hers of t he similar t ype.
² For one t rial, t he dist ribut ion above is called t he Bernoulli dist ribut ion. The mean and t he variance is as follows.
E(X j ) = 0 ¢q + 1 ¢p = p
V (X j )=E X ¡ (E [X ]) = 0 ¢q + 1 ¢p ¡p = p(1 ¡ p)
5.4.2 D i st r i busi B i nomi al
² The number of successes in n Bernoulli t rials is a random variable, X . ² What is t he probabilit y t hat m out of n t rials ar e success?
m p n¡ m
m;n =p q
² There are
n!
x!(n ¡ x)!
² So t he t ot al probability of m successes out of n t r ials is given by
p x q n¡ x x = 0; 1; 2; : : : ; n
² Mean and variance: consider t he binomial dist ribut ion as t he sum of n independent Bernoulli t rials. Thus
E [X ] = p + p + : : : + p = np
V (X ) = pq + pq + : : : + pq = npq ² Example 6.10 on page 198
5.4.3 D i st r i busi Geom et r i k
² The number of t rials needed in a Bernoulli t r ial t o achieve t he …rst success is a random variable t hat follows t he geomet ric dist ribut ion.
² The dist ribut ion is given by
½ q x¡1 p x = 1; 2; : : : p(x) =
0 ot her wi se
² The mean is calculat ed as follows.
because t he sequence converges, so we can exchange t he order of sum- mat ion and t he di¤erent iat ion.
² Example 6.11 on page 199
5.4.4 D i st r i busi Poi sson
The Poisson dist ribut ion has t he following pdf
8 < e ¡® ® x x = 0; 1; : : :
p(x) =
x!
0 other wise
wher e ® > 0 ² Poisson dist ribut ion propert y: mean and var iance are t he same
E(X ) = ® = V(X )
² The cdf of t he Poisson dist ribut ion
because ® is a const ant and t he increase rat e of i ! is much fast er t han t hat of ® i , t he dist ribut ion is most ly det ermined by t he … rst a few it ems.
² Poisson dist ribut ion is most useful in Poisson processwhich isused most
frequent ly in describing such phenomena as t elephone call arrivals. ² Example 6.12, 6.13, 6.14 on page 200, page 201