CHAPTER-3
LITY bisTRI8UTIONS
DiscRETE RANDOM UARIA8LES AD PROBABIL
RANDoM VARIABLE
fur experinent , a narton variabl
Sample space s 3ame in
Cng hule that associates with eoch outcome
a no -.
S
In mathemahuo a random variable s a unction
wose 'omain is the samble sace and whose ranges L
Bet o real number
BERNOULI RAN DOM UARIABLE
variable whose ony Bossible value is o1
ny andorn
Types of Randon Variable
cONINUOUS
biscrete
It is an RV whase Bossibl e beth ef the folicing pb'y
value either cohstitute a. ) Iis set of þossible ualues
Arite et or else can be Corsüst either of al! no.S in
isted i an ininite 3touence Single interual oY the no
in which #here is a first dine or all number n disjotnt
urior
element, Jecond element and
2) oRV has the þrob.
P(X =c) = 0
’ PROBAgII TY DbISTRI8UTioN FoR DISCRE TE RV
hokability di:tibut ion of say KSays how the total bsolbabi
i'y f i is distYibvted aan(adocated tu) the arious
’ PROBAB!LITY MASS FUNCTION
MF of a dizcreie fV is aened jor evey nember x by b n)
"frobabiliby
PARAMETER OF A PRORAAILUTY OSTRIBUTION.
that can be
"PARAMETER: Suppose p() depent
on a
qvantiy
asigned any one of a mumber of posSIble valu el., tR
en ch digere nt valual" determinng a ayferent forobabilih
disiributior. Such a is called a barameter
quamtig
distibu tion
FAMILY : Tae collaction of all phcbabily distriovtio for dfere
nt values of the farameter is caled a
distibutio
CUMULA TIE DISTRI8UTloN FUNCTIONCCbF)
The cDF fC) a discrete RV ariable x wi th þmf p(1)
is degined Jor eveg
Cor any number , Fn) s the brob. hat the obsered
Value X will be at moSt X.
Step Fuun ction
V,
for X a discrete
Josh of F() will G.
have a mp at everg
possible value of X and
D-2
Valu co .
Juch a arop is calod a tep unchon lo
GxPEC-TE D VA LUE S:
let X be a discrete, RV wtth a set of possible values Dand
PME (),
h expe ded vatue Y meun value of X, denoted
EC) or Ux or U
nED
HEAVY TAIL
Anyy distributionn having a darge
u.
amoun t
prob. or from
Exbectad ualue for a functionn hn)
’ RULES OF
EXP6CTEO VALUE
E(ax +h) a.Ecn) +b
where a,b are Constant
UARIANL6 OF X
let have X Pave prof p )
Vor
K denoted by U() or
exþeted alue u. hun the
o or o-2
UC)
E((-)2)
Sandard Deviation
X( SD )
FoR MULA
’ RuLE5 OF =Ea)-Eg
UAKIA NCE
V(axo)
:lal.
-V(ax) a² Vor (x)
BiNO MIAL PRORARLITY DISTRI8UTIoN
) Experinent consist ot a dequence
Call ed tnials, here n is ixed n
n smaler experiment
aduance expejmnent
2)Each thial
Can hesult úw Diu ed the same two 0SStble Qutomel
wich are
3) The trial are
denoted y by SucLes (5) ond ailure -
independent So tRat the autome or
Barticutar tiat does not nytuen e the outome en õny otker
trial.
4) Tu ob. of sUccess P(s)
de noted By P. constant fsom tial to trial)
7smple sjpace (no trial) n is at most
Srze, the eperiment can be 57. Bopdotho
a binomial analyze as thro wgh t ware
exacty
’Mean
experimert.
Eca)= hp p=Succes
Variance Varcx) = hpy J 1Ja ure
-’ n no trials
PolSSON PRo&ABILITY
DISTRIBUT ION
p(:u) = e
*
Vartance Var () =
aves tho hole al an event
times (K)
witRin guven hotpterval
n oning certbin number
time 08 sþace.
CHAPTER-4
COrINUOUS RANDOM VAKIA8 LE AND PRORARILI Tå
DISTR8UTIoN
fROBABILEY DENSITY FUNTLON (P0F)
The pdfca) is d
Junch on fn) ch Kaat for ny
DENJSITY UKE
Crapn of f(n)
a b
’ UNIFORM DISTRIBUTI DN
9 otßerwi Se
’ cumulative istibution clunction (CDE)
FL) Jor a continuou8 k is delned or euery nymels e
xis condi nuo us RV wik rot fon) and cE f ) then af
evry Xot which th deivative f'a) enist, FO) 4c1)
’ EXPELTED VAWE
continvoU 8 Ux with PDE f() as
ès any funehion f
- hlx)
E[ntnJD - n h(a).4a) d
VARIANCE
- Var (x) = (*-u). fen)ax
’ NoRMAL DISTRIBUT|ON on normal distributiorn
A continuo Us RVX is soid to haue
with arameter u and o where )
W2n
STAN DARD NO{M AL DSTR1B VTJON
Tme narmal distribubOw wit Barameter values u=o an d
the standard norma) distibutíon .
a stand ard non mal Sandom
andom variable Rng
Vaable and wil
PDE of z is
f(Zio,1):L.
-2
cle noted b
Shaced area
Standard normal (2)
2
normal dist&? bution WR u and stondard
9f- X has a
deviahen then
ar arbit rary normal distribution (100p) hereRle
(0oopJ tor stondard normal
Ajproximatsl hu binom ial distribvion,
Let x be a binomial RV based ohu a trlals With sUccess
brobabilit p. Then if tne binomial þrobabiliy kisgrom
is not too skewed, x has apaoX a normal distibU ion
for 2 - a possble value oX
PCx<) = B(, nyp) farea under norna
Curve to
2t.5 -np
