S9) Joo odebendent RV x¢y aue wuiformly dis ibutes rn the

Kame ihterval o,) . eind P(X-Y zo)

Soln:. P(X-Y)

let z X-y P(Z0)

z X+(-)

fz()=x(z) fy -9)
frla)

tyl9)

‘fz()
So, P(yo) > Y.
S: -l
P(X-Y>o) -
Analysis:x-r o).
P(X-Y> o) P(x>y)
=P(x Is the large st)

Note:
X Jhree
Random Vaiables fomly dis ibuted i
the same ileval Hen forobabi Lily that
P(X, is he largest) =a
 ldeuically diehi bulid
Let x, Xa and3 be independent amd O].Jhe
vauibles with f m dsnbuon
robatbi lily hat p{xxa} is -

Son:.

Use
C.R-v.
dK) 13
= P(X1tX-X}so)
=k
X+X1<K
k =0S
K àk= O.49 to osS

P(X1+X, <X3)=dK

X}+X >X3.
P ( 2 X3) (-*)ak. Xj+X, >K Prob. of these
Aevente to g t t

i - | -t7) aneoucty
Slmlt
 Pick 4 PicK A2
P( Xio ls largestb =

4 chantes.
b/w o ei wth equal
* 4 an be qny valo Nhold cu simultancously
ocuu, 10 events
Fo t i ev ent to
A10 K

Xg <k

ç dK.
x Prob of XIo a K undej K ie K.
legs han K is aea
that X, LÀ (rmlt'þicaton doe
events to Ocu simutaneouly Aine rdeendent
Hence for a l
OCUra nce)
P( X1olslargest) =

OREMou uiln n
x CENTRAL LINI|T THE
Vauable are adaed hen
STATEMENT:. Sf Large number of Randomn
densily tuntin t reswuing uncion aboroa ches to
GAUSSIAN DtNSITY AUNC TION.

Analyeu X= X + X +X3t X4 t Gaussan

PDA
Note. For
eabe Anayex t t h o .
ndiideas
PDE aw taken
unifom. Jhese
PDE taken Cank
any pDE.

Sosne
Paratc
shoe
 Dsael ROrCm Vai able
* Jhe txla) c o aOo be onpube tx (cussir He Ce cLT
that case the Regutant PDE weU n o be
f l s tos DRV. PDE
Cntiuoey hen Resetant E wu be
* But when the RV is
GAUSSIAN.
is t au aLpe
x Cenbal Linit Jheore mis wefui m Noise
the
" Cs xmi tfet thhough
no. ot Nose Emls aue added as
the nannel. aclodtive n
NQise u R o m n n naue and hey co
NOe R)aue aldes to
nature. Hence when ape no 4 cosneA Qut to be Giauis ian
reguta PDE
the channel t
+nl) nt = ntn, +ng
s
St) Rx

NOISE
Note
ww be dla ferened with iterent
Jranumì tedt Simal
AGNAS at Vauious Baces n nature sothat
Densilg
wiuI be additive PDF.
¥NOSe nle ference AlOise abraches to GAUSSIAN
func of Reewant
RV:
ae Gaussim
also Taussin RV
Fnen (Ax tBy)ue
me Corstamt
whre AeB cue

*Analcis :
and tx () a giren Continuoy
eProvded X Ry Cue boh
dx Random Vaiable.
then ty (H) = fx(a)
dis biuted m -T, to . Find fyly) Auch that
Cs wi fornly
Y tan X ?
So1n.
O other wise.
 lasste
Date
Page

y a

dx

Thn
 X Y Y= tan x
E ytanz

then, da
tys)

yl) y:)| X ’

Given that
fxta) = iua+s) -uz-s)}
CRV

and,

2-5

hen' fyl?
Son:.
ux+s) - u(x

Since taking ny a Þossible

Ye{o, I} Vales. Hence RV.
valueg
Xt k e a heuge Range ot
Randorn vaiable.
X is Continou
Hence
 P(Y=) P(-2-s <x<2-5) > ½.
Ply=o) = P(X <-2-5)+ P(x>2-5) %.
aleeseeys the Demiy
Dewsily Functn ot DRV res lz m brnpubes ·" Hence
Her
*
functi m may be qven m. Pyly)>sl9) + sly-)|

X* à booped ttoouh Hat wave

Vaiable.
an RV
G)4 O mean Guesidensl y funch of Reeultine Rnoln
Rectif r . Pind
HwR

‘fr(2)
y Gion that e
fx(z)

Now
Y=x xo

HR

wth ig U) a hat
‘fyt)
tlore :. Jhe orcblern
zhe qrëa ¢ hdensity tunceet heie
be
AkouLd Come.O t to rnpube
yty Hence addillonal yo
to be incudod at
tignal SY) has
Snce HwR
4OCi a l w -Ve
 Hence the rersentallon fyC9) wiw be:
fyl9) stu) + J e

y<o

DIFFERENTJAL EN TROPY
ç CaLled a Dsferen
Randon Vaiable
* Ehoby ot oninu Qus
tial Enrot
vaiable is jven as:
* En3oby of bisCYete Random

HC)= - )loga Rr( f

Random Vaiable s en n!
¥ EnlaopH ot ContinOus

H() [fua) lega)d

Rondomn Vaiiable us uni fony disbuted in (o,a).
AConinous Random Vaiable. close to a
robaoun
Enliopy of is
ind difterential x
tk(), Pob. a! hds high
lg)
ajni an it
SOIn:. fxl) H‘
(ighuncer H .uancEatai niy)
(lees
Ya
a
Note :.
*for CRV tHe tcx) wll e
-0
manium with unitm dish
a
buin as
Po Eawal Probahiity
Jt og, a dx
forobaa, ul is equal Hene
Llog,a x a H(x) is manm
a

HO)=log, a bts/ Snbol En oop of CRV wiy be ma m.

 Gien that
e

PSU's

+hen 7differes hal Entroy s

x i s CRV.
PSu'S
also
HOX)- logaf2nerf logT f Janote

x Given that (x-a)2

L
x= N(a,o?)- fxla) Janga e

then
loga fYne (osa)t
Y= a x + b ; & o
Q)aiven that
hen shaw that
HCY)= H(X)+log.a
ao

son
Y= ax+b

(fala) leg a adx)

fx(n)
So,
> ( )log,adx a

-po

HLY) = log, a + Jfx(2) log, ’H(V) = H(x) + log, a.

 oney
Note :

f X4y are ndependent Radom vaiable hen:.

E[xY] = Elx] ECyJ(7hs
E[x']; E[xJELY]
CO-VARIANcE. ECKyECED
Vadable on qnother.
* t speç fies dependenne ef one Random
Matheroaicaiy My= Expetati a x.
Cov{x, y}: E[(x-Mz)(Y-M)] My> expectation of y.

Variance [x]: MS§[x-]

-E[X-Ma)?]

Cov{ X,y}- E[X-M) (Y-My)I

- EDxY]-My E[x]-MyetY]+ MzMy
> ELxY]- MyMy
[xY]-ELx]ECY]
cov$x,yt = E
are &ld to be ON CORRELA TED Or
9f Covx,y = O : thon x ¢ y
ORTHo GoNAL
ndependent;thex
ECY]
cov{x, yt =o ECxY]- Ex]
D., but ls
Xp y ane Sdependent hen wçu be UN CO RRELATE
Convee i not fue.

Given that
then Variane Lz] ?
 0090'
So, Elz] = aE[x] t bELy]
Mz
Varlance Cz]= E [[z- Mz)?]

- [(ax+by)--tam, b,j]
=E|{a(x-M)+ blY-My)}|
=E[a(x-M)'+ b(Y-My) 2abK-M)(y - M)]
-aE[Cx-M)?] +bE[(y- My)?]
+ 2ab E[-Mx) (y-My)]

+ 2b Cov [x, Y]
VAR [Z]= aYar [x] +b²var[y]
covx, y] =0.
X Y are ndependenty then
then, VARC1
VARCz]= avAR Cx]+ b2

Note:
Y a X
VAR[y]= a²VARLX]
VARL]= a' VAR LX7

FONC TION Fxy(X,y)7

TOINJ CUMULATIVE DIS TRI BUTIO N
OR
JOINT DISTRIBUTION FUNC TION E L Fxy(,y)]

Mathe mataly Poob. of x taking valee

Fy (x, ) P(X<x; Y<y) upto
Va l e whto

Fxy (oo, o) P(X<oo, so)

 Exy(X,o) P(xsx; yso) Fy( -o, y) = o

= Fx x) Exy(*,-) =o.

Fv(a, ) P(xso, y<Y)

a)
Fry (-oo) = P[xs-o; ys -

Fxy(-,-)>
-JDint Densil function.
Fxy(Z,n) -

SStey(a, y)d ady =)

SSfxy(x, 9)dMdy -Fy)

Fxy (oy)

Note:. Densi'g fune

be calewlalid toom Joint
xBndiidal doily tunch can Nte.
P[XJ = I P(X, 9K).
K

A kcalled as
MARHNAL
DENSITY FUNC TIONS.
fy (9) = Jixy (a, 9)dx

+ry.) )
 9) A Joint Demily nn is qiien by.
fxy (*,y) = xye-ty
otherw ise.

(x+y)
fxy (X9) - ye u(*) uly)
Funcans of xey.
i) Aind Marginal Densily nal Dewily tunChon ?
( ) caled as Condiio
i1) Pind fx

SoD:. i) #a)= J fxy (7,)dy

S*ye uc*)uy) dy
f2ye-( / )dag u () uty)dy
yla
ze ucy ye uy) dy

uu) (ye
Let = t

ydy = dt
dt

u() | e d t
2

= xe fx(a) Xe ula).

2e 2u) e
SiilaryUy).
y l9) ye
 * Hence it look Ke .

fky (a,v) fx(a) fyly). 4Joint Densil, funcn equal to

mutibiaton of rdividal
Densi y tunticon. Hence scid
So, xPy are ndep endent that x 44 ae ndeperdent
Fromrfomnalon
ii) y("y) fxy (zy tr(). ty(y) Jheoy :
P(AY) = P(XY)
ty 9) fyl4) P()A

>fy (a) P H(%) = H(x)l when Xy

Jnlepen
x () H(Y%) - H(y dent
lectee-9g
Stochasttc
guesing
meaus
KANDOM ROCESS (STOCHASTIC PROLESs)
RAN DOM
tuncon ot tine is called as
* Random Vai able
PROCESS. temp at 4 PM)
temþ at 3 PM Y
X

Coeate a RV bA
iy we hue t
Jo reasure he temp: of and e celectuon ot eaG exþt
¥ Lape expt each da
Conducting
caled as
SAMPLE SPACE. predicL
A
we ant to ALEYe he tem. a cly at
supþose unegs we ae
t3PM, then cue an't oedct
pcn
awaiu able c - a to gues outidr
anase tho behawiar ot tenb at t=3pm t a
*Hence to and colcc
o ee fem a seies ot eopeumcnls athered duing
AtI hese dala
we
l ewey day
the ten. f tho caled SA PLE hd e þlae Lohee tha, qre
c
erpeæ Ment pace we comþlot
Kebt ç cald SAMPE SPAE Frcn Sq mle the tem.
a cubo
he enaig . o n and c n oY edict m
day if we otbserve he vaialn o tem. and Csns ondig
e wiu qet oo no ot Randn Vaiabla
GAM t0 GP uncno Hrne is Randm Process.
Samle space. Jenu RV aas a
Rande m QandomVaable
(Samjpleng of
RandomnProcessRV).
cUrTTme 9AM ults in
at Res
Renom SsrgameE Sample func !
Romdon oiess
x ndesrine
o amount Ot
RV's.
MGAM xtaampe
9AM hpm GPM we get
Pt of Ki
Day 2 ample tun^2
6PM
GRI,
Ath samplePE

Samþle func 7
6pM.
caled ENSEMBLE func|
Colcin ot au samþlc funch ia Sn Ensemble anoiSample
same
ENSEMBLE 1we have o
CAM PLE SPACE points
large
Vaable we need to hae
no.

about Jem
Jo bettes coro nnt
-Samples funchon. Prorees.
RY called as Rnd
D0. ot
Casllection of uncountabk u ales Rondn bocess.
finduon of tine
yA0, RV a a
Random vawables a calod a
no.Of
iolection of uncountable

Proess resls m Rand on vaiables on

Gompoing f Rard om OR RAN DOM
called as ENSEMBE
Collcdion o.Cample Funch s
PROULCs.
 *STATISTICAL AVE RAGES OF RANDOM PROCESS .
cPayamete
xJhey aue af two types.
) Ensem ble Averases.Cau colection of Sample tune eKen joto Considetin
b) Jime Averaes. to calculali Hhem).
(ny one Sanupie funchn a Constdered).
Note!.
* Ensem ble Avg- caleulated by indig average of U Sample point
at one putiula tiane.
yo Ensemble Avg u pOssi ble
*Jlne Average u caleulalod for Jsgmple tunch over a time
duraton ofal the s0mþle pointS.
Note:

Statiatical Averapes at a xed tione Arstant over a qaoup ot

Sanlc Puncani a aled as ENSEMALE AVERAGES -
'Siaital Averages iren Sanple 4nch calculalsd over a
sheu fic bme nlei val u caled as TIME AVERAGEs.
Ensemble Aveages qres beter resus of vaiiaticns
ene we gnerd caewa enermble
*ENSGMBLE AVERAGES!. Aveasges ony but not Jime Aerages

a) MEAN!.
-denity funco
at Yed ime
anstant ie OAM.

* Jhe S a l e a token foorn Samle tunaon at same

time t," à cale as Romon vauab anl
Same space t m where we an þlo the densi teuncn x.
b) MSO!:

c) VARI ANCE:.
) AUTDCORRE LA TION
UNeTION:Grtagrg Corszimilet
desree of
sìmilqilj between
ren Rand om PoOCeLs
Averge of ’=e[x(t)- xts)]. at tUD dH- i'mne
rodue
Sampkes. 4x) desyee of
e.
< Similan l dt t

Prdutof X, +x of s anpe
tunc I.
tz
Comment
to. be Considered to better stonces.
* A U e Sample tunch are aat wo dUf. time
abo degree of Siùlany H-Hme hstnt
# Produt of S a n p l e savg
. o fs r t a D
aue calealed anal then gh degsee of similany.
calcuaod ia igh then less degree of Snilang
Calcuted en
i) Sf av itf
ProCess at tuo
Note Silanl of a Random
XACE A meesure of
oetanes.

t ta-tj -l<time difterente.

asoalad as!.
t¡-t,=z i) Sear chins Parameter.
iü) Scannins Pa ameter.
Eatt=(attt)
bef wuich h e hee Searching
x we aye seas chi ng t
for siilaij
pauamtes z.
XJIME AVERAGES!

i) MEAN. t

JxLHat
Kx> = m, = ta-tr
t

ii) MSO :.
-xHdt
ta-ti ty

iü) VARIANCE:.
mg-o,
PROCESS!.
*CLASSIFICATION OF RANDOM
:
i) ERGODIC RAN DOM PROLESs averages then i t is
aleA
u equal to hne
Averares
* Sf Ensem ble
RGODIC RAN DOM PROCESS.
as
tostant shold be same
at evey time
) Ehsem ble AYg. besqme.
ewey Sample tunch should
ii) Jimc Avg to should be e5uat to Jime
AVG
ble Av
ii) Reslti ng Ensem cPaticay).
PROCESS doegn t exist rn Reailg
x CRrODIC RANDoM RAN DOM PROCESS:.
ü) STRICT SENSE S TATIONARY CSSS)
PROCESS !
STATIONARY RAN DOM
oR STAS
ages aue dependent of Hoe then caled as
x f StallsHCal Ave (Ensemble Averaset only)
PROCCSS.
SsS RAN DONM
ble Averares ae Aame at each and everg
oWeg stanb ime nstant
x ensem reyese o not rue. au sss
s S SS RP but the
y Every ERGODIC RP Coneliion ot ERGODC RAN DOM PossIALE.
RP -saifiei ony the J<f
almast not possible.
t SSS RANDOM PROCESS LA
 2n5emlile
) wIDE SENSE STATION ARY CWSS) RAN DOM PROCESS.
u said to be wss roided MEAN ACF are
A Randon PDcess
ndebendent of ime:
Hrne mean that MEAN
i)Mean Ahould be todependentof Snstat
&huld be sarme at evay Hme
e ri) ACA should be dependcnt of tHme
the Revece is oot
WSS but
SsS Rardon Poces wiu be shoud be ndependent
*Eve
true as n SssRP au ststical avg:
otHne. since it is
m, but not m(
Fos SSRP the mean is
rdebendent of time

ACF sheld be an cticn of < onyie

exLH)- xlt+] bndependent
* f t.
(z=a-b)
t) ex lb)Tahauld be tunction of t-t," ony
AcE- EDxL AcE shold be tuncn
at Ii one
Hence Lt cam be Concuded that
dfference only kIsSRP Sn ce
tunc af (t +t) hen it is not
4f Aee s
t+ty = 2t+t.
satifpn ondh of
Henre ACF iT func oft" bence not

et)

3 4 (o 94

ACrobtained b/ w 3 to 9 shold be sqme as that obfunes

and 94 to I00. Hen Ce uue Com Say that ACA
fo 4 to lo
time.
C Drncles endent af fferente ACE sbould be Aame.
th Dame Hme
That istu
 Note :.
ie with
XACF Ah ould be tuncon af time dfference
sam time diffevence ACE should be oame.
where A wo
x ) = A Corut+ )
ARandom Process uniw y is
Co q 7 endon vaiahie
au ctostants and serate
buted m (o, 2)·
gnd ACA ?
a) Find Mean ?
KISS RP or not Trme.
b) Veity
Soln:.

time.
Randon vaiable but as a tunch
Randorn Proce ss is
ast vaies hence vaçes uncdeterminstially wth egal
en oto QT
9 vcuies betue
tine instant t ;
chanes.

= ELx+)]= E[ACoT (wot +)]

a) Mean
|4 Cos (wot+). flode

A Cos( wot+)d9

A &in (0ot+0 &in (wot tan) -Snwott

E [XLt)] - o ’ m, ndependent of time).

b) Aef E[xLt)x (t)]

t, +9)|
=EACos (wo t,+)· ACos (wo
=EA{Cos [woltG)+291 + Cos [ wo lE, ti) ] |
A Cos (wot,+wo t, +20)fl) d9 +ooslty-t)} fle)eie.
 Cos (wstúst,+20) 40 +
A'osfws (tg-ti)} JtO)de
kSSRP.
denoti on for

ACE
Jrrne.
mean t ACE are on deþendent of
a kIssRP&le Rxlt, ) .
¥ Hence it denote cF as
Romdon Poocess we de
¥ Fos nornal
+)?
Repeat above With, x)> ASin (wot
(9) de.
Soln-a) E X)] = A&n(wotte)

E[xLt)]=0
ELxlt)-xlE)]
b) ACF =
t ) Ain (wo ts + ) |
&n(wot,
= E A tG+ost; t 20)|
2
Cos (wo ty - lw,ti)- Cos (coo
EL aslwwat, +39)1
- e[A os (wot, -Woti) -

a WsSRp
u asO
* Hente i t
Note 4x+)
eCxH- x(t+)]= E[xH. x(t-z)]|
for wsSRP.
(t-) t (t+)
Hence
Rrcz) > e[x+) - x(t+)7 Rx(z) > Rxl-z)
ECxL+)- x(t-)]

* Por Wss RP, AcE a even unctin of T.

Process is giren by x -
X)- ACor
ACos ( wot + D) she re Af wo
(wbt+D)
&s) A Random unifenly ddisbuted n (o,n).
aue Ctnstant. 9 is Random Viable
KSSRP or not ?
veity
SOn.

a) Mean EDxL)]= ACox (wot +e)- lo)dd

Sin(wot+e)
&n (wot +)- &n ( bot+o)
Smwot
-A&n oot - A

ExL+)]> -2A_&n vot , +) |<-NOT AKWSSRP.

IT
 oont A ç a Random
a
A RP o iren fonly
by x4)- ACos wot
eial betel n o, ). d mean+ ACr
Vutable wea

Son. xH= ACorwst

)Mean:.
E[xt)] - ACoswtflAldA CosDat- ELA]

fAlaswst d LCoswot

CosustfAdA
Cosust
LCosbt

Corws tg ·ECA]
Ca)dA.
orwot,: Cos wst, x A^
= Cor wg t, - Cosw,t,x 4|
ACE = Cort, Corwt,

Hence it is not a WsSRP.

y
 t}> Ao a Conatant
Cos(ot + ) .
x
) A Rand con Process ç qien bd ç unfon y disibute
AeO are dependent Random vauiables . A
ibued rin (o, a ) . nd nean
iFornly da
9 ç
m (V, V) andxt)?
and ACC of fle)

-V

* 9f x¢ au mdependent then
E[xy]* Ex]. ELY]

i) Mean = E[x+)] E[A Coswot +)

ACosfwot, + }1
E[xtt) -xtts)] E [ACos fwot, tof.
i) Ace wot, +9)
> E A o (wot,+0) os (
los (wot, +0)]
=t [A]- E[Cos(wot, +9)

(av 3

So, CA'] E[ (os (wot, +9) - Coso (wot, +)] { Rejor Guec}

Hence it is WssRP.
 Rantom
Cshetant;dom
A +I Pocess
Kane are
B
ero ronean todeÙendentqiven by.
and vaçance
A COs wGt
Radom vasables+BSncoot . wois
Sojn. eah
E[A] = ECr] =o
then
So,
MS^= Variqne
avia?
Mean + ACF havi
of

ECA]= ¬CAJ- ECa]

O.
) Mean:.

e CACoxwst7 t ECB Srà wot7

onw, t ECA1+ Snwot ¬CAI

So, |EDx )]-o

ii) ACE -) e[xlt) x Lt)I

+ Si agswot ELABJ+S wot, Sinwo t, ECA]

Cos ws t, Coswot E CAI + &n wot,Shwot eCA]

T Coswo t,(exwota t Sin coot, in wo ta?

2 Cos wolt-t)

Rx (T)
kJSSRP.
Y HenCe