InfotltltlUoD.,conwncd in lIl:irs work ba-sbee-n obtained by Tata McCh::aw":H~ll, from SOu.re~s:OO:~i.evedID be ndihlble. However, neither Taut. .McOmw ..

HiU nOr its ;authorsgnanultee me accum.c:y 'Or com,pleteness of any lafonmation publi.shed he:r.eill~ and ne.ilther lata M.co'mw-um nor H~ authorsshaU be res;pons[b1e for any errors" omissioas, or damages ms:ing Gut of use oflh~,s i.uotmiiJUou. This wo.rk. ISpUblisbed with the under,5Cmding that 'Tata.McGraw .. Ri.n and i.ls auUm.rs are sup,plying InJotmaJtioD 'but meDal aUumpting to render ,en,gbll.ooring OF other proJic,ssional 8ervi,c;cs. ,~'f .such. .services

"-.;I .:.. • - i' .. - c.' ]

are reqjlJhw.,,~!ne, aSS1SInUlce 0_ .an a:pp,oopn.ij!tt-pr(heS5U()!Il:a

s.bouLdbe ,sought

No .pan of ttli;!! pubJ1i<:il,Uon can, be reproducoo in anY~()mlor by any means

"til - 'h- •. • - ""f'Ib' 'bl"sl1-

. - ,: ' ,I - - I' I" 'i ,- " 'I - ,-'-, - ,I \ -I" ',_ - '--, _-, "II '111- ;-, _ -,' 'I'" ' , I' -', c~ -, ,,- , '1-':- -~ -,-

WI OliU t .. e p.no! w.rn.uen permUismn 0 . . e pu t. .ers

- - -

ElciVCRm reprint,20Ot7 RALLCiRXYRCVVX

'This ,edido.n can be exported from Indla only by the,publisnernt 'Tata .McGraw-Bin PubUsmnl Compoo_J' Limited

P1u'blisb~b~' 'rata McO[a.w~HjUPUblisbi·ng~Cbrn_pany U mtted,

7 We-it Patel Nagar~ NewDeIM ll'O 008. 'typeset at Script. M;ali:c-rs,

19 A-l~B DDA.'Market, Paschlm VIhar, New Delhi ~u)063 .a:ndpri.1lted at RaS.htriya Printers~M;o 135 Pane.b.sheel Garden~Na.veeB. Shai1c1am~ Deltu. 1110032

The McGm w·HifI Companies - ~ .. ' .~r

Copyriphted materia'

Preface to tlie Second lJditio'1l

I

Random Experiment J

M;a:tbema:~ica:~ or AP'fiori. Definjt:i(m of ProbabUit,y 1 Smtis,uc.a1. or Apo;s~er,IO!dDefinU:iio:n .of Pro'b.ab:i~ i'9' .2' Axiomatic Defmition ofProbabUity ,2' Conditiona,1Poobabmty 4

[ndepegdeo,tE'vents 5

Theor-em offotal Probability U}

Ba1J',e'~'''',' ·TIf- ...... rem or The 10' eem .' of P ·~b· abil ,'t- ~". "",,·f·C.· ..,-' ,,,~, c, . 'L '" Uo.;u ....... JJ "'", .... ' _,r", ... _. ....._~:yRchl y v~ '. ·uiU",es

B,emouJU ~s Trials 2.1

De Moivre=LaEdace Approxima:Uon 1(]

Generalisation ofBemonllPs 'Theorem Mubmmnial Di stJributjoJ]

19

26

:2:,,_RandQ~m_%u:iabl;Wi

DmscreteLlandJom liaRable . .16. Probability,Funct:ion 3," CmrlffiHons..Random_¥ruiahl e "Z Probabil.tly Dens,itj Functkm 37' Cwnula,tiveDiskibudo!nF'unc;tion (cliO ,3'B "lioperties ofti;e odfF(x), 38'

Special Di.strilbutionsJ8

D~s.cfet;e Distributi (ms ~'9' CorntmuouSsDistributions :JJ2 Iwo-Dimemslon:al R:andom Variables 5:8 ProbabUity F-uncti()11 or(x~ ,}') 59

JOint Probability Densit)rFuDclicm 59 CumulatJ've D istribuHanFufldi on 5'9. Poopenies of .F'():,. ]1) 6() Marginal.P.r'ObabUi.ty Dhitribution (JO

X Contents:

Conditional P£iobabilUy Disu:U:mdoUl. 61 mndependent RVs 6.1

Random Vec,l)O[s 62

Margin al Probability .D:i:uribmdQJ:lI nIX ~ f i •. p,· , :Margimal Pr,obabUUy Dis·mbutj,oo af ~r :{ J~P*j ~

.. .

6ti 66

.3". Fl!lrnftinns ,of Randlm Variables

EunctioQnfOne Radom llariaMe B2 OlIlc(OlJlilct1nm ""Two Random Variables 90 Twofiurructimls, of'Two R,andow, Vnriahl,es 93

4. Strdis:ti.cal.A ve'rage-s: 120

Expected V.aluc$ ura Two .. D]:lllC:"S~onai~ RV J 2,1

Pro, ~~i.es of Ex,- ~ -,ilied VaI,- as 1.22'

___ pe~,~ ~- _ _ __ UC_ ~ __

Orrvarj'ancearnd OGue~~atjnn 123

P· J,.,_····.3.: .• •

.rope.ftles _ ~

Co.ndluona~ 'Elil.pec:t~d Values ,124' Properties, 1.25

Cbamcterist itc-J!uncUo,fJ 141

.

PI',opcmi:es of :MOF' 141

,PtopeJJltics ,of 'Characteristic Function ,142 C'omrlJllan't GeneradngFuD,CUCUIJ {C,OF} J'44 1'olnt Cb:aFfI c~ed s~lj,c FuncUnlUl.l44

Bounds, o:nPrdbabili:tie s .1:55 Tcb.ebychefflnequaUty 1 ss

Bi,e'naymets lllieq.tlO~jtY ,1 $16

Sc:hw.adz :lmeglUallUy 157, Cauc.hJy@,s'(:hwartz .bleq,a:~jtJ ,15'1

C'onvergellC'e: Ccmcep,ts and Cen!ralLimit Tbeorem 164 C'e.nna], L1mh TneClfieRl (L~apornJno:frs,P~)rm)i 1,05 Cel3!u"al.Umit T.bcort'Al. (Li.ndberg~Le:vy":s :fiorm) ,1'65

is., Sum!: SpedaIPr,obabi~ityDistr:i.uti.o,ns.'18 ltrtmdluctioni I tR

,Sp@cbd :DisiCreiteDistrib,UOl1ts 178

,Mean and Variance ·oft'he Blno[lilial OtstributJIon .1.i~9 Recnrre:neeFormrulllafnrr UlI,e C,rmttall MnmeFl'~S of :tbeJBinom ;.11 Dilliltribl~d.cm 180

Po,isSlo:n DlsttUmUcfll ,as 'LimUing FOili.m ofB,il1l.0mia'i .DlstributiOlb 1I1.J :Meanand Variance of Poisson Dis:mbution 183

R,ecu11lence Fonnlda fo:r the Central M():~ents of thePois'SlonDls:tribu.tJion. .183

'M'('lHlmilUd Vadance of (]eoUlle;tdc Disb:ibutlcu:ru {'8:5

M,ean and Variance, of Hyperj:eome"1.dc Disldbul'io,n llUi

CopyJligh!ed mataria'

Contents- &I.

BiDomial DisitribuUoVJJ as LimitingP'orm ,of Bypet,geomemc DismbuUDP 188' rSpec;lal Conti:ntlOU5 Dlstribultions 2JO

Mome:l1J.~s of !the Unifonn. Dls,tribuUon U Cd. 0) .2J 1 ~M~MOOd V:~dmo~ Q:ftb~exp(l<1t;entim 'Distribution

2}'2'

M·e·mand Variance of Er.lam.g Di:s,triiblllti:on 214 ReprroducUve; Property rQf (laguna Dismbutio:[! 214

Relation be~veeD.tbe: Dlstdbuuam Functions, (cdr) of the Edq D.isuibu.tionwith A~' l~or Simple 'Gamma Dismribl!lUon~

and. 1(.POiS5iOO Distributwon)' 2.15

DensiryFlllcuon offue W,eibuU Dis,mbQluon ,216 Mean and Variance: of WeU]!1!iI11: Dis;tribIllUo1!l .216

Standard Nnrmal Distribution 217 NQfm~Pro'baJbimy Curve 21,8

P,rope.rtm,es o:f'me Normal D.istrllimdonN (u. (1) 218 ]mponance ofNonrud Disilributio:n 2'27

i., Tests of Hmotheses 263

Parameters and 5Ua;~hstics. 263

Sampling Distribution .263 Estilnath:m :8fld Te~tinl of HYPQitbeses

Tests of H:lporth.eses aD,a Tests of Signifi,cance 264 CriUca] Ba:lJglon amd 'Levei~ of Sig.trllificUJoo 265 :Errors: hm. Testing IJif Hypotheses, 266

One-Tailed and 'Two-TIUed Tests 266

C:rltica~ Values, or Signific.ani: VabJJ:e~ 2'66 Procedure for1ic5ting ofHypothes~s ,2.67 ln~e['val.Bstimadon ofF'o,ubtiorn Pammereli:5 268 'Tests of'signifilcmmce far :I.Mgt!: SamJl!~es 268

TestIS of,significlAcefor Small ,s,amp!~es 29'.1 's~udEmt~ s .t~DifS·tri~lDJt;on 291

Pfopeni.es of t~Di,smbu.tion 2'9.'2

Uses Ole t;.cDis,tribut]on 292'

Cdth:al VallUJes o,f r and the '~'Tabl,e 2SU! Snedooot~5J'£'- Dk;cribuuon 2'2;)' Prol!)erUes. Oil" the F .. Distributiorn 2.96 U~e ·of F~DiRtriblllfion 296

F~tesl 296

Cbi .. Square" Dis_tribution 311 Preperties ofX2 .. Dlstrib:utiom 311 Uses ,@f .l:2~-Dj;strilbuUon 3.11

Xl .. Test of Oood!necss of Fh 3]2 Conditions for the Validity of i' 2-Test 31.2 .l~t Test 'of ludepende!nce of AU db utes 3.12'

Copyrighled materia

Random Viariables D

I:i{x. x . x ).

It,.,. '." .. '.' .').... . .' !'P;P]',

,J'~ I ,Xz't .13'· ,:: '..... .. ('._ ". .,. ') .. :

. Iv ~... x .. , x~ .' Ja:~: ..... ), ,",' .J',

1 .. '1.' .. · WO. ftUI:'1h"D1"'.., .... rnT'" Il'(:-;-C) ~.'I·

.......................................................... -Hn '-- .-·~~,LC~B' IH-~I ~

ExAMPLE, I

'Three balls aredrawn at ,random, 'Whhoutreplace~me:nt from a. box. oo:tl),taiDing2 whltei, 3! red aDd '* b~:ack balls. If 1l denotes, the DlJJm!be:r o.f whltebaJlls drawa and Y dellGtesthenumber ,of red balls, dtawBt find th,ej(!l"int. poo!bability distribution lof (Xt ,Y).

As there are only 2 w1hltebaUs, in the box,]{!can take the values 0, 1 and 2 and "Y " - '! • ... ke * ~ V'''~iles 10- 1- 2,-" ·d,' ·3-

caD. WAi_ wd3 ... oiUU. __ .. '~ _, _ aD_ '."

P(X: O~ ,r= 0) := P(ckaw.ing .3 baDs none ofwhicb is white lor red)

=P('all the 3 balls drawn are b~ack)

-, c-319-'· 1

!=4C'C:,>=~ _·3 21

P(X: O,Y,~ 1) ,: p(dmwing 1 red Wid 2, blac-k. baUs,)

3!C~ x4Ci .3

'~L :'~

'9!G 14

.' . . _.- . . ... '. . . -'. ' .. ' ,~3 c2x 4 qt. ..,' . ,. '. .. '. 1

S.unilarl.y, P(X = 0, r .... 2) ii::i: '. ..... '. ~, .';' P(X .... · 0" Y ii:i. 3) .... - -

9.(.'3 784-

P~ '(- v ,_ '11-y,r- ,_ ,0·) - !. P-'-(!V_ t ~yc - 1- ') - 2., p·(·x~~_ 1. 'r· ,_, 2-') _ . 1.

<4'~ I" "~ '~'_'_ r ~ ~i':-': '\,4'1, ,~, lit",. ,.., .s r« ==-,1 ",I." 'iiii· J'I!' l' 'iiii" ~'-- - __ 'II

- . .' 7'. '. . 7· . 14

P(X~: 1, .f = 3) = 0 (since only :3 balls ;ate drawn)

PiX' , 2· . 'U' - OC )" . 1 ... p .. ·C""V' . "' .... f" . t )., . I ,,01 .. 'v.,. Y' C 2· .. ")., n. '.

','=: "it .l' '=.' = -, .; -;-:JA =:,L'1! ", =, .li.,I,=, __ .:~ i--\A;;; ~I .~:-.:;;: ._', ;;1(1;,

- '2: 1 ... . - 28· .'

P{X == 2, r == 3) ;:: 0

The joinJtprobability distri:bution o:f !(X,y) may be ,represented in the form af a ta!ule as ,given, below:

x y
!O I 2
, .
0 1 3 1.
21 ];4 '7
i :I, :2 l
1. 1
!
Ii 7 i 14
, .
2 1 1 0
-- .
21 :2& 3

o

0.

Copyr'ghted materia'

.y, Probability" Statisucs:andRaDd.omProc.eues

SII1II vIall t/Jsrt:lJpm,hn.b.iiiJiu ~ ",

-i---------~, 2

Forthe bivariate pJiObability distribution Q,f (,X', Y) gilen below"find P(X'S 1.).

P(Y <' 3-)' Dl'Vtl'" 1-, r-r- ..... 1,]') p(',"v e' I/'Y'''''' 3') P(' 'Y' -.L' 3/'X-- .... ' 1·)···· . d P·(X.' Y' ,.,.. AI)

.', ' ~ :" Ii, \,A, ~ ,_ ~ 31, ,.~,~" .... ~'."",;;;:z. '" ~. _. an_ :"',_'c + ," :=:., '"t."

Ix'Z 1 :2 3 4 I 5 6
I 0 0 0 1132 2132 I 2/,32 3132
I l llilU~ : :I.lUi liS , ~1I8 U8 118
2; 1132 I 113:2 i UM I 11M 0 2/:64,
'-'.' -'''' P(XS;: I) = PfX= 0) <t' ,P(X; 1)

6 6

= I P(X:~ 0, y =11 -+ ,I,P(X 1= 1, Y=j)

j-I .J~~

_ ,(""".J C , 1 2, 2 " 3), _ ( .... ] ,1 iI, '. ,._ 1): :; 0+0+,',',-+', _'+~,'+-, + ,,-, ,,''+' ........ , ",_ +~+ +---t ..... ' .

, 32 32 3-2 32,1.6 U;;, :8 8 :8, 8 "

1 5 7

=..:..+~;:::;:~

.til 8 :8

PfY:S 3) =; p(¥'= 1) + P(1 =:2) -+ P(Y:= 3)

2 3

:.0: L P(}l';: it, y'=, I) +1: P(X'= l, y= 2)

2

.. I ,P(X := it, ,r;: 3),

~-I

.. (O+l~+ 3~)+(h m~ +3~)+(~'+ ~+~)

3 3, 11 23,

:=;-, ,,+,-+-:;:;:;:-

,32 . 32 64 64

l 3

P:(X s: 1" Y S 3)::.; I ,P(X;;; o~ Y' ,= j) -+ I,p(x;;: 1. Y =: )1

j:=:!ll,

Copyr'ghlcd materia'

Raadom. Variables §i

(0'" ' n' 1)" 1(' l , 1 , I.)'" 9

=:+vT32,+ :16 + l6+-g;32

P(X< llY -e 3) = p(X.~ t y ~3) = . 9/32 .,= 18,

, " _"" , _ 'P(Y S 3) :23.1 642,3

P(Y< 31X< I)~ .. P(XSl, YS3) _ .9/3,2, _ _!_

,,-.. ,," .... _ P('XS1) _. 7/8 _ 2.8

432

P{X + f S 4) = L- P(X :::;: Otf :=;) '+ L 'p,(X == t, y= J) +l: PtX =2~, y= j)

i""l, i""l - 1",t

V"""""-'-,''J!I,''

~L.r.o, i:t

Thejoiat probability mass function of (X,Y) is given by p~~ y)= k(2l;+ 3,), x;;:; O~ I, 2; y= 1 ~ 2, 3.Pind ,all the mar,gillaJand coadi:uoaal probabilitydi:smbu.nons. Also :findthe; probaibiU.ty disldbutiOD, of (X'T Y).

The- jOlB,t probability distribution o.f(Xtt)is, gl.vm. below. Th,erelcV'ant prob-abHUies have beep ,compu.ted by Q~sip~g thegj,ven.law.

1

2

!j

x

I

3

o

6k:

9k

5k

8k

llk

1- 2

71

13k

3 2

I . ,~I P(,Xi!~ y)) = ] j=1 ,I =0

i.e., thesum of ruT the: probaibilitiesmme table is equal to 1. i.e •. ~ 72k ..... 1.,

Copyriphted rnmeria'

16.. If j(x, y) ~ k(~ ..... x .... y), 0 -c x~ y< . ~. ~ is a joint deBsityf;lInc:tioD~ :indk. 17.. If the joint pdf ,of (X~ 1') .isl(x~ .)'I)'::,!" 0 :s; Xt y :$;. 2~ find .P(X ... rs 1), .. 18.. If thejOlint pdf of (X,. y) isl(;t~ ,) .... '6e-~3.f~ x ~ OJ '1 L~ Otfind tbem.a:rgiiDat

density of.X ,andcondjijanaJ density o:f ,f given, X.

19.. 'The jpd'f of (X" .f) In. given byJ{x,y)= e-{z + Y}, I{) S x,, )" -e 00 •. Ale X and. Y imdependeDlt?Wlly1

20. Define a random v,ecror wiman example.,

21... Defllleilhejom.t density and dlsuibutioo i\m.ctiOM of,an.n..mm.ensio:natRV.

H.ow aretheyre]ated?

Pad.i8

22. If.X' denote~s;tbe, DUmber of aces and. ythe,uwnberofqo.een.sobmined. when. 2" cardsere drawn ,at random '(Wiithout .replacement) from. a deck. o.f cards" obtain the joint pro'balbrllh.y dbtdblltioR of (X~ Y).

23 ... The jOint ptubability function of two discrete RVsX and Y is given by c'(x ") :::: cl") ~- -.,,), , w;'b. ',--, •. ' ,- .-ilc - .' '.,. - me .. 111• ! ....... '.- , '. sum that '0' S xS; F.-~Y --'\.""" + Y t , ..• ere.% aDuy can. assUi __ WiJm~gers --.------- '., -.'~

2 and 0 :$; y S3,. andJ(x, y)=O otherwise, ('a) findtBevalue of c and 1(b~1 find P{X ~ ltY ~,2). (MKU ~ '. Apr. 96)

2-4., De join.t. probabil:Uy distrlbu!U.OD. o:f 81. two-dimensional di.s.crete RV (X" Y)I i:s given ibeiow ~

x
y I I~
0 l 2 3 4
.0 0 0.01 , O,.OB O~O5 0.07
i
n I O.O,I O.(}2 II 0.04 O~o.S 0.06
~,
2 00'1 0.03 il 005 0.05 O.OS
•• L I ". :',1,'" ~',:
3 0 . .0:1 0.02 II 004- 0.06 0.00
L~i _'-'.' '.' 0.09 I

0.08 I

0.06

0.05

(i) FiBd .P(X> 1') aDd .P{max (_X, 11 ;:::;1 3) and

(ii)find lhep.roba.biUfy· distdbutioD Qf the RV Z ,; min (X'"Yl

2-5. The, input to abbwy eommunicadon :sysreml, deaored bya RVX~ tates on one of two values 0 or I wlt.bprobabUUies :3/4 3001/4 :tflspecuvely. 'Because: Qf errors caused bYDC)jse in tbeisystem:, the Oltput, .fdiffen,from the IA!p,ut: occas:ionaUy . The; behaviour .of the. eomraunicauon syslte.m is modeled by tbeconditioBalprobalhUities giv1en below:

P(Y~. ilK . .::. 1) = 3/4· and P(Y ~ Olx.~ '11) '5i 7{8

Find 0) .P(Y;; U. (ii.)P(Y ;;; 0) aad on) .P(X;;; 111;;· n,

26. The following: tabterepr"8seo'ts tile jOlmt probalbWl.y djsb:ibution. of the discrete RV (X't Y). Find an the mM!,iMl and conditi.aBa1 distributions ..

Copyr'ghted materia'

Dff~erentiaili;ng (1,) wilth respectto ~ we get

~.IJl ... ~;;;;. ~. ""II...t ~--k(r-Ai""l e~' At + (r e~;t l!""l-e"".t,.t') (r-l~.ll:~].·

..:I" ~ .. 111:··· '. .'. .

'1.1 "" r =. 0 'i

I. ~."'.... ._:l.·tl. ~,-A '.': .. ,.r-.' ~I 1(·· .. ,·._.1 .. \ ... , .. "+1

:= - i f1t-:11 ,+ L ','" Jl. 4,1

,"~I .r.

t. Lll 1(· '~J:"

I.c". 1J1, .'1 """' 1I!Io.. . -+ ,t. J'l1.1

.JI.'i!;.' -, dJt. r_" .

U:s:ingrec~uaeiBcJe relation (2)~ we .mm-y compute moments. of n.i.ghe,r 'or-der, provided we Down m'om,6Rt of lower order.

PUlliD.1t k =: 1 in (2)~ we get

P2 .. 1 (~1 .. Po )

= 1 (./ 14J; 1 and .u'~; 0)

(2)

PuU:ing.k =:2, in (2)~ we get

. .( dp2 '.: ).1 :.

113 ~ l- - dl,~ + 2p:, ,= A.

Puttin..g k I;;. 310 {2)~ we get

P'''~{ ~; + 3p, ) = A(3A. + I)

I'JJe ~:nt~mlillg .prop.~rtJ 01101J' J>Qi.sr.(w {li~t:riblltj,()f.~ ii;;J tot: 891,all!J~1 itoS !JIi!alrJ~ tlan.4#ce·Olitt thiril~or:llet' ti.f,ttfl'al mOl/JirJllt.

:DeClnition: Let. the RV X denole the number of tria1s of ,8 ramdom expc-ri.meo,t .requiredt.a obtaintb,e fitst. success (oeeutrenee of , an eV1ent.A) .. Obviously X can assume tbe vabaes I J' 2,~ 3 ~ ....

Norw X' ~ 1'" if and. only if the first, (.r ~ n trials: result in. :failure (O<o'Currence o.f .A ) and. me rib trialresuits in sUC<:-'e'8S (oocurrence: of A). Rence

P LX =r)= q~lp; ,.; I, 2,~ 3~,. ,. "11 0::1

'·W· '-- ....... '." JIIS.I.~

P~A··'" ..... P··'·' an d P fi)1 = 11

- \\; I - .- u . \1"1" "1"

If .X is: a discrete RV that can assume the values 1 ~ 2! 3J ., •. ,.~ .r;:oo saehthat its pnJbaibiU,ty mass, 11mcUonis gi:ven by

P(X = r) = q,.,...l! p; r = 1~2~ .... ". IX> where p +q :: I them X is said to follow age,o.metrfc dill;,'rib'ution.

Copyriphted materia'

1_,

~- ,.'.' ,11 , II '_: ~_

= pet T' q + + ... =)

P'

=, ···-···:=:t

l-q

'M' ..

:L P(X ~ r) '~'L q'- P

n2 '3" ]

:=,Plt,1 +, "-~q + ,.~,q~ .+, ,!"!I,,., + QQ,

=» (I '-lJr1; .'~'

p

41'

-

_- '~{".1 ., 1 ).. . .'} ~l

.... P ,~ .. ' , \,r + .' - r ~.' ,q

,.'~'l

if2- '('} \-3, t 'Ii .)'~2]1

= lh_, ' .. ~ . .q,- ~ V,- ,q> ._

(2 lJ' 1 Co I

=p....... .- ~ '-'. =:: (2·, '-P-.)';; rn + n.,.~ .

•. :5 "J '2 \.. .' .' '2 \, - ,,;0

" p p.' p- p

'V' ' .. ..t. 'v) E·--·(·X·l\ 1- :rl"'('X)' . '}:2

·,,~",ai\;"',Jl :; " .-. I,,~_, -} - :,~ rll,,-, .. \'~

.samet/mel the /J'fO,Y'fJoilityl!Jm:r fiUt£tiO'lJ ola ,IJ()1fJdri& RV X i;S' lak,tII1 as

n,/V = ..,1 = or ft" .. = ,Ill l' 2' . '00' , ... ,L~m 'h ~ n= t

J.""'lm'.r~· - L~I 71' . Lr.~ T tit· , 'J' I. I. Iii Ji - - . ~~- ,_. Jr ~~' I

Copyriphted materia'

m: Prnbabiilit;y, ,statistics and Random Processes

.. 1lyJJ8tJ.~_ .... - -, --··--_~IO, h~r..~Nft-,

.... ' .. ,. ., , .. r reo,11K1CTHi '.101,11'""'''''''''

'Lo ':-'-' - - -- -"_

If X represents the number of defecdves::found" whenn items are drawn, widJout :replacementfi:um, ,a. lot of .N ite,mscon'tainlng k defeetlvesand (N - ,I), nondefectiv~s"clearlJ

icC ·(N'····~k)C,·.'· .

, ... ,-r \'- - J,(n-,.r) __ .... _,_ ... .

D(X.·, ~, 't·), ~ . . ···r -0· 11 2··. . -'·i'il'(n. :,.)

,r' .. "._, -- .' ',' '!"'!'!"!' !Ii . ~ _, ,Ij"" _'~ '~'~'IIII'. ,~w_~ _."",' .. ,,1\ '

. NC~ , . - "

,If,U > ,I; #Jm 1m l1JtJxlmNm tldJlJtJ o/x is~' 11 n «; k~,lhlm' the IIIml1l1lmm;/Nl q/,Xi.t",

i.t.j flit mflXi!1l171'1l Mille ufX ill min ,(n" k)"

i.I'J r ,tdN IQ:b; tm' WJ11I'JO" :1~ 2" •• O} min (n,. k).

DefieUlo.n,: 'If X is a illscfere~ RV that can assume" non-negadve values 0", l, 2" ". "., such that. itsprobabili,ty massfuuctlon is given by

_. ... .kC"'(N -k}C{li~'i'L .. . .. ' ... ." ._.

P(X;g 1')::::1 . . ; r= Ot' 1~2".".~R1lD fn,~ k)· ..

- - Ne - ... '\.

--1'1

tben X is said to ,[oUaw ,a' .nypergeometric ,distribu.tiO'n~ wi,th theparame:ters N~ Ie

8odn.

_ __ _ _ _ tIJ In th:ep.m.babi'Jity .tfJw-s fottmo/J ofX~ r ~4N Ut· aJs«me:d It; .take the I

i i mINts O~ _,1 J' . 2~ .' , .• ,11, wh!dJ /J .tnlt w!Jtnn< k, HIII' .11!0('# n >- .k, .r't1:tI l:pk~ tlMt¥iI:fNt$ OJ . 1,~ 2t- •• k. In ,(Jth'1'11JOnll~

P(X ~ .. r) := OJ ~ht!J r § k ~ t,.k + 2~ .. "'J s: This lI'nhte (IJn11f~&zt'ro) 0/ Ihe,p,rolmbllity iX,p,rotiaea ~. tlJe'ptV.ba,{,jlitJ lltaSJfit:'fllll})/.t jormlrlfJ'itself, till,ce kef ~ O,.for I: ,. ~. A: + 1, A:f2~ ... , ,t. I:

Thill in' the; ,mine ,ofP,(X ,= r) min tn,. ,k;) (Q-:n' berepla(ea .I?J' N.

(2) Hipe'l:.e(J:NJ~ITi(. ,di'srnblltiotl i;1 a le.tl1imatJt. probalJiJil§ distribHlioflj ilr1:fe

~f,'(,'r'{N'" k)C(jI'_' r) .",.PI(X ~. 1')' ~. L,

,.~,Q T~(.l' .1' .... l Cit

(1)

1

~. . NC~ ~. t since

.NC

_. - _!/

,rl'

L, kef!' (1J'.~. k) C_,. ; t'DeJficiel# ,oj,x1j in (1 + x/ (1 + x)N-1

= «M_jji(ientoj," in (1 + x)N - NC

-",

J,!,eaa an'd.Vartaace ·Qf Hypergeoim,etric .D.istribution E~'V\ ~ ~. X P

""no.1 -.-'-' '-'r-r

1I

~ I r ke,. ... (N - k}Ct~)/NC:n

,-0

Copyriphted materia'

(on putting 1(= i- .I and. r ~'r - 1) == ,~!. (N _. ne, _ '1' '~by-- step- U) .ita.bote (2) aiven above]

NC .. ·..· . ,n t - --.' . . et .

;,:- . "n:

n,l

':="""""--=

N

EtX2) ;: .E{X(X ~ I) + .1]

E'{V(V ~')} ,nk

:;;;;; Y':A'4, - (: + . -=-_-

N

.~ n '1:: .... ' + ~.. r(r- 1). kC . ,(N- k)C.-, . ~ . .1 He

N ,L,' . . T- .. ' - ~ ,_ r) ,~

.'. ,:::,0

n,l: k(k-l)r.I. . .. , ." ,'.'

~ - ... .- ~ .(1.= 2;'C ' ... ,. IN· k)C

.... N NC .. ·.,. ~,'.- --'}·(r-2)· \··'-····'tn-t)

- - -. "~r'-I2:

--, 1,("". 'l')n'-'l

n.k A ,1\, - ,: ~,- .......•. ..'.' . ."

::::ii. N.:'.' .,+.. N C: . .~_ VCr" (N .... 2 ... k')C~1.I ~2 .~. Ip

.' . 'II! 1"'''' B .

(onpu:ttmg ,K= Ie .~. 2: and. " == ,'.~ 2) n k _ ,1(1: -- 0-,,_,

..... ,_ ... ". '(N,~ 2)C cOl,.

N-- N'··C··iPI"""·

, .. "-.~

[by :5te,p 0) in DOle (2)ghen above] n k . .k{k - n n(n.-1.)

:;:::'------= 'T' '- .'" '. .

IV' N(N,~l)

Var{X) ..... E{X2)- {E(X»)2

,n It .k(k ~ l.)-n(n .~.l) ,n:2 ,ltl

~:-+, ..... ,. '

- .II - ,HeN-I) N2

,; N2;:-1) [N(N-l) .. N(Jt - 1)(11-1) - (N- 1)11;1;]

,~ n.' I.'.' . . ['N2:- Nn _ ~Ik +-~';1I

_ -'l _ _ _. - . '~ .. ' - ~lh -- ""',J

.N"I(N -1)

~n k(N·~ k) (.N- n) N2(N'-Q'

Copyriphted materia'

21Bl Probability, S:ttaisti.c.s and .Random .. Processes

---------·.ExAMPLE U) ---------

The mean o.n:mld:ng s~[eng,tn of (he cables suppUed by a manufaenieer is 1800~. w.ithan SD' af 100 .. 8)' anew tecih!nique in the B'1aI.llfactl!JriDg proce:ss,. it is ,claimed '[balthe tlreati:ng s,t:re.ng.tb of (be cabie has Iacreased; 'fo test' this c~aim" .3 sample of' 50 cables is tested and it is fuund that the. mean. breaking strength is 1850.. Can we support the claim at 1% LOS?

HereP = 1850~ n ~ SO~, P := 1:800 ander » 100. Ho: X = p.

H!~' .X:> p ..

One-tailed (.ri.gbt.~talled) test. is to be used. Le(.LOS .. ;::::; I %. The[efore~ ,lo;= 2, .. 33.

z~x-lJ, := 1850-1~800 ::; 3 .. 54

. al-r;; ~OOIJ50

c. Il~>Z,~

Therefore" the ,difference bet'weJenx and Jl~8 :signincaDJt .a(: l % ]e:ve,[,. Le., .80 is m~,ected anoMI lsaecepted ..

T.h.lJ:t: is". buca. (InJ Ihe sample daaa, wc;ltia.y .support the claim of increase in blea'Klillg stFeng,dl~.

,---------, 'DI_ .... ..-'1t? III" •.. --------~c~l",

"DJ;e mean value of a random sample of 60 items W11S fOllndto be l45,t with an SD ,of40. Find the '9.5% co.nfide:noe limits for tbe IK"Pulali.on mean. Wha!t size of the sample, is teq,urred In c4J:umate the; pOllu~.aUom mean wi.thin :; ,of its actual value whll. 9.5% 'Of more ccOufideoce:r usimg the :samp,le mean?

Since tb,epopuJaUon SD ais not given, we CaD .app,roximat'e it by the samp.1,e

SDs. Therd'ore, 95% confidence limits lor Jl are given by 11!1 s 1.96.

8./11'

i.e ..

'-11 n6 s <# ....~·"·I (itC S

x ~. 11, .•. 7 ..... r::.~".~..::I;.fJ ~,x + ·.jv -r .... -~'

'lin .yn

"

I.:e .•

11 ..... ~ _ l.'9(5.xAO ~'" , ... ,'. .( 11 A C - t9~)( 4l()

~~J 46fJ _I,.. - .1'-.;1 + Joo

134.9 S;.}).: S;' l55.I

We have to findth,e value of n. suchlhall: .P tx~.5 S P S x+ .S·} ~ 0·9.5 i.e, P {-5 S # .~ .. X' ,5.l ~. 0,195:

Le, P{ 1I,It - .i II s :5} 2; 0.95 or P' Ui- fJ. ~ s s) ~ 0-95

Copyriphted materia'