Scribd
Upload
179 views27 pages

PSLP Notes 1

Uploaded by

aza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
179 views27 pages

PSLP Notes 1

Uploaded by

aza
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 27
Dr.Nishy Gupte Poobabi iby anol Statistical Moclels : The wonsl “probable” means * Likely” oF mest Sih ely -to be-true’, % “Likely, though mot codaby to occur’, Tn other toorsls, We have Om uncertain situation, Whine We, Camot Predict in advance, lohat is forgo happen. The probability -theovy provioles a Mothynatical, moclel to stuoly the uncertam situations , This course gives us the Rrotbeclye of how -to Wanipulote ‘Probability in orcltr-to assess Aishs ancl make bebfesclecision Unteughout ScitvdiFic Omsl ARO olisciplines. Sable Spaces and Events ! Ranslom Experiment! An exptrimerd that can Acsult In clifferec outcomes, even though it is Apeated in the Same mann Weytime, is calleol a Adnolom exferimtct, Scumple Space! The set of all passible outcomes of @ Aanclom experiment is called He sample space of the experiment, The Samble shace is lmoted as 8, A sample space is often cLefined based om the objectives of the analysis. The fo ting excmple illustrates severot alternatives ; * Exomble Mollded Pfastic fast 2 ©) Considler am ex-pertrannt In which we select raokeled plastic part and measure its thickness thew sample space for its -Hickness might be Se R= PxX[x 70h (+ Sthichness cannot be -/ve) ©) TF we consider the thickness betes [9 deel Il milimedes, thee Sof xligcx és wn > Oo @ (Ava) Nc cane)! DA. Nishy Gupta Following Laws for see operoHons are Useful herei eS) & (E)'2 E i AUB =BUA, ANB= BNA UID Distributive Laws : (Ausg)nc¢ =(Anc)UC BNC) awl (ANB) UC= (AUC) N (Buc) Cy _Detigan's fav (Aug)'= Ang! and (ANB)'= Aug Counting Techwiquis: Semitimen it is not easy. to olefermune the Ui f Cudcomts in UtntS, Tn that Case, we ese the relat Farle ant other Cotding technig uss that Cnably oneto easily dketermine the rinbe of gtdlcomis na Sdumfle spac. or eveads And this, in-twn, aWlews probabilities of events to be oletermined, Mulfliation Rule (for counting techniguss) Assume an optradion Can be dlescAibeol as a squtnce of R steps, amt — the no. of ous of completing step lism, Amol the v0. of " a n Sth Qis no tor Cece Of compbeting stepl vey nol the ma. ef w af compiling step 3 is m3 tes cach of Comp betings stead, Ano Soforth , Ten, The total mumbue of 204s of amcplating “He optrttion i = 1, xXNgX-- -X My Dr Nishy Gubte © Subpase that Wwe wonkto furddase q tablel computer, We can choose cithtr a Longe or small scrtens q@ 64GB, 12868, % 250 GB stevage Capacity. dmol a black or white cover , Hota mont differnt options do we have ? Bithe rubliplication hubs, the-total number of possible. ObHons = Ox3xa =a A Example { Sol Permutations: Arothte useful caleudation is the number of ordtred. Sequumee, op the dlintds of a set, A permutation Of the Dumtets is aw ovelercel Seq ence, of the Ulbmtnds. For ex: abe, Ach, bac, bea, cal anol cha Qe CW of tho -pvrudations of the abanent Wf fyb, The ‘Menbtw of “permutations of 1 difforit Ulmeds ig n! Whine MI = Wx(m=1)x(m-a) x-- Xaxl THis Aasult Folllaas faem “Hy rutiplication Ault , Examble: Th hee, studeds poaticibate in a Ane Combetition. Bo h ow mawy Ways Com-the First thee Prizes be won j Comnot why more than One. phize ? Sol BY tHe mulpDeabin Aue -the no. ef emis ~— = Sl = 3xaxl= 6 AU $4 student + Th some situstens, we are interesteol in “the number Of only Soma of the Utntats of a see, In thor Fesubt(which ols Follow s-from -the omulFiph Of Mrteptinas Cane he Folly cation rule) ts useful! Dr Nishu Gupte @ Permutations of Subsets ! The tuber of -pomudotions of subsets of eineds selocted roma sub gf differnd eDumerk Is =P, = PP = Mx Cra) X Oneal ~~ x(n-RT) = Grant Exanmble: Ty the above exOmple if Ten stuoltnts spantiupate ina AML Competition. Then the number of woasys tar First three Prizer (ig a studet Commat tory rare than One prize ) = Wp 2 lol = loxde8 = Tao a CH also FoLllurss farm reall cte sol) * Somatimas we ade IN eresteal in counting. the vuunbte of ordered Stqutmee, fer obyects that ane not all oliffiud, In thot cae Following postu Is Useful: The number of permutations of 1 = M-bMgt— = +My objects of which mn, one of one type, nny are. Of Second type, ---) Anol My Oe Of eh tye = mt MI ml ---mg1 Example: (Bar Codes) A -paxt is fobeled by Print ng ith four thick Lines, three meolium Linas, dad -too thin bnes. TH ead, ordering, of the wine Linw rrbreseats @ olifferadt Jobe, hors many diffent Lobuls cam be Gertrcted by using “this scheme’ Zh Nishy Gupta (6) Cor Sat: By the above Aasult, the number of assible chant (obels 4 ys FM 2 xB eT Px _ hago sr gee ~ Combinations: Another counting problim of inforast isthe nUmber of subsets of fr ebinints that cam be seltctedl from @ seh-ap rn Cldmtads, Here, ordir ts not impontart, These are callled combinations, TRe numbur Ci combinations, subsets f sizer, thot com be seltthek From a set of m clumads, is denot eal ~ (2) oa C 1 "0, nd, (3) . Che "O.= a Al (n-a)| Example! fl pranteal ancuit boarcl has eight oiffenunt Lncastons in whith a Compared Can be Placesl Te -Five ioletical Combonety One to be Placed onthe board, how ray diffennd desians On possibLe ¢ Sol Each design is a subset of size five “From the eigdt Locations that Ore to contarn the componets. “The numbev of possible dsigns = 8, = 8! ee X, 513] = 56 A Dh Nishu Gupta @® ‘TrienPrrtations anal Axloms of. Probability Probability. is useot to quantity the Likelihood, or chance, thot Ow outcome Of & Admdom exptrimed wil) occur , “The chance of Aan today is 30%," is a statement tho qudatities ow Feeling about Hho ‘possibility of hain, The. Likelihood of an outcome is quandified be assigning. a number From the infOwed (0) 1]to -the outcome Cor a percentage Frome 0 -t0 lov',) * The -faobabl of. wm outcome. can be interpacted as gue Subjective -probahibility, on degree of bebef, that the cufame will occur, * Probabilite, For a Aamalom ex pehimed Gre oF ben assigned on the basis °F A Reasomrable model of the systers undir study One approach isto base probability Assignments on “the. Simple concept of equally, Dkely outcomes . Probability of each, equally Likely outcome! Whenever @ sarnple space consists of N possi ble outcomes thot arn equally Likaby, the ‘probability Of. each outcome ts YN, Probability ef an event: For a discrite semble spose, the ‘probability of am evedt Ey demoted a, PCE), equals the sur of the probabilitie, of the oudcmms in E, Dri Nishe Gupta. (fo) Combining “the abwe-two, we get the Follouaing oLefintion * Whenewe, 4 sample space consists of v possible ovcome, thot ore ruddy exclusive anol equally Sikes Anol, Mm Of them ane favourahle to On ever Ey thew P(E)= Favorable no-of outcome, bFaveer of E Total no. of outcomes = (Mathematical or classical olopiniti y of probabiity ) Anathe: int ‘ation robabibity is baseal on the conceptual moolel of nepeading replications of the Adudom &xbumect, Tin n tAlals Cexprrimets), an eved E habpens m timer, thin te “probability of happening of E is give by. PCE) = Lim 2 (Stabistical ox Empirical cleft Nan nN OF Probability ) TF PCE) =0 ,-then ib indlicote, the outcome E will not Occur and P(E) =1 indicate; the oufcome E wilh Occur certainly. Note: Da ish Cuple © Eaneble’ A ros done Of Womted Coos piiuty (none fod} tort prebabdan, oF, 6 0-3, 0:5 aod Oot, Aaspeettody, lef A dterte ths Und {0,4}, B4te Ged Foot db, and C tke wtJ'd) The. of ths cubtewes PIAS O1407-0y PIB = b3965+b) = 09 Plol= ot POA) = 0-8 Pel) = 0-1 Ceo) =0.9 Ate, Angeshl, $0 PLAN) -03 AUB~ fo,bre,d},s0 PCAUB) =0.140.3405+0-1=] and ANCe 4 se PLAN) = 9 ExQuple: Supfese 0 balch Conains six parts fa,bye,d,e-F3 Omol-tun ports are selrceal Aandemly, withouk Aeplocemtat, Suppose that fat F is duficive , buf “the others one g-oodl, What is Hy Probabibty thay fat -f abpurs in thy sample ? SA Her the sample space consists of all possible Cuwordlted) ‘Pours sclrtecl without Acblacemect, Total mo. of crdtomts = So, = 6! _ oxS_ IS ayy 8 lee E dnote the Quad “that Pod -F is in He Sample, thi Eo Plath, PRL fot), fat, fern J P(E) = 5 (0 ch gutcome is equally Liksly) a al DANishuGupte (2) Axioms of Probability: The axtems clon ot cletermine -probabilitie, Howevn, te axloms enable us to easily caleulate the “probabibitio of some Uitats from knowledge of the sprobabilitin Of oH euents, Axioms! Probability js @ number that is assigned -to each mumber of & collfection of ty tads Proms fandom ex penuct that satistiv the Following. “paopectin 0 EFS isthe sample spac and E is On ued in a Admolone experimunt , GO) Pts)= 1 () O PCE,UE,) = PCE:)+ PCE) * The above axioms imply the Following Aaseels ! tw PC$)=0 ce) PCE!) =1-PCE) Co) Ff-the evted E} is containtal in “the vent Ex bev ESE, thin P(E) < PCEs) Pacey (@) '/ SU¢=S9 P(SU$)= Pls) PCSI+ P(p)= PCs) > Pld) =0 Profs mem) () EUE'=S =) PCEIHPCE) =PlS) (by axion ) SJ PCEDPCE)= 1 (by anions ti)” Pe! )=1-PCE) proves! / DaNishy Gupta cs) o Eic Ey pe] (eines) Nei=¢ ond (E\NEa)UEr= Eo. eo By axlom (3), PCELNE,) + PCE: =P CEs) > PCEINEs) = PLEa)— PCE) by axiom (a), PLEINE) 70 > PlE)—P(E:) z 0 > PLE) < P CEs) Provesl Addition Rules: Unions of events, suchas AUB; inctersectionap eyerks, Uh ANB; Anal comblemuds of events SUCh a Ar Ont commenty of infucse, Probobili Quan! The probability of Aor B ity inttafuted % PCAUB) anal is given by the Qololition rule a | PCAUB) = PCA) + P(B)—PCANB) TFA and & ane mutually exclusive tds, thew PCAUB)= pcA) + PCB) Note! Tn care of terre evinbs, PCAUBUC)= PCA)+ PCB)-+ PCc)— PCANB)—PCANC)—PIBNC) + PCANBDC) anol if A,B ond C ow ruts ly exchusive events the PCAUBUC) = PCA)+ P¢R) + Pec) / Da Nishy Gupte (iy) / Note that a collection Of Weds Ey, Ex ,--1E, is saicl to be rnudualy exclusive if for off Poins Exn Ey=¢ ud For a collection of ‘muckuolly, exclusive events, PCEVUELU--UE;,) = PCE) + P(Ex)+--+ P(Ex) Example: Disks of polycarbonate plaste from q supplitn are zeck. for scratch and shock nasistamce. The reseutts Fro 100 disks are summarized as -folllas | Shoda AgSisttnee rg fet senateh pusistance | d™ To q Vow 1g 5 (a) TF a disk is seleded at handom tohat is the ‘probability that its scratch hesistance is high Ono) its shock Aesistance. is Wi CL) TE a disk is seeded ot Aamolem, lolot isthe probabil sthod its scnalch fasistamee is Nigh oy its shock Aesislance is high 2 Considlt-tha eve thot a disk has Nias, scrates Arsistamce nel “tht Cvindthat a disk har Wal, shock resistance . Are these tio autos mrshaally exclusive ? Sof Consicdlte the Following Wents ' E> scaagch agsistonce is Wigh for a disk Fg —Setsete of (c) Ez; Shock ” high, n ny Te foo gives date! _Dh- Nishy Gupte. a3) | q E,NEs) =To <9, / loo 7 £ (PCE UEy = -ecernEs) ++ pee) P(E!) =(—O . / eee and (EUE,)= E/N Es Wee Take E=EUEsm@© = 0-95 A i Co +s pCE\n Es) = 0.7 #0 SC Thetwo weds are not rnustuably exclisive, Example TP A, Band C one rauctclly exclusive eutnts “itt, PCA) =0-2, P(B)203 Guol PlC)= 04, determine the Following probabilite,' (a) PCAUgUc) cb) PCANBNC) co ecans) -() eccaus)nc] ce) pCa'ng’nc') Sol Ca) P(AUBUC) = PCA)+ PCB) + PCC) by additie, Aule ar A,B, Core exclusive vents ) = 0atds +04 = 09 A tb) PCANBNC)= PCH) =0 L CAs, cone chat] (@ prangi=PCH=0 Bb Cn , , 4 (4) P[AU8)NC}= PL~ancywanel c , Ly = P[ pug] = PCH) =0 Cel P(Ainginc’) - PcAvBuC)! - 1—pcAvsuc) = l-ot=al A Cb ) Dh Nishy Gupte (© /Erantle Tha aty 3 catty remspaper X,Y dno % are -publisheel, ‘oho of the pevble of the city Aeadl X, Sof, Asad Y, 264 Akaol Z amst dove Avad both X duel ¥, 15%, Ataol X dud Z Ons 10%, read Yond ZX and 249, reno! all the three Papers . Caloulade the percentage of peaple who donot fad ony of the thee papers. Sol Let ts cbefine the events A The nest that people Ataal * earanee G32» 1 Coon Give PCA) = He =04 , PB) = B= 0S , Plc) = 200.2 a” yo R 4 PANE) = 20-02, PCANC)= ney P(BNC) = Bao ao Pranenc)=a-4- 9.084 leo _ We watt final PCANBNC), Pp canenc)= PCauBuc)! = 1—PCAUBUC) = | -[ Peale Pra)+Pce)— PCANB)—PCANC)-PBNC) + PCANBNC)] = l-[04+05¢6.2-02-oely—Gl+ 0024] = 10-674 = 0-326 "Bal ]o People oto not Acaol ony of the thse fafers, Dr. Nishy Gubta @ Conditional Poobability! The Probanility of am evect B undtr the Rnouledage thod the otucome toil be In event A is clenat col We PCB/A) and this is called the concliticnal prebability of B Jivin A, Note! Gnditienal probabilition can be calculated clined ly ,Thex Probabilities cans alse be dettrmincol-from the formal olefinetion OF conditional Probability: _ The conditional probability Of aw evted B givin On wed A, demoted ar pc6/), is P(B/A) = PCANB)/ PCA For P(A) > 9, This cefrrition Can be Undtistcad in a special cart in whith aff outcomes of & AQudem expurimed are equally Likely, wf thor are v total outcome, P(A) = number of cutcomes inA n Also, Pca NB) = numbte of cudcomes in ANE ™ PCAN 8)/ PCA) = Tuber f cubcome im ANB inh ey 4 outa in A PCB/A) can be indtrphtted a, ths Atfadive fFregusmcy rd Cutnt B Canons Ha triads that produce an customs in tung A Phi Nish Gupte (7) Random Samples! To select rdnclomly implies that at each step of the Sample, the ems thot rvmam inthe batch are equally Jikwly to be selectest Muthlcatin ot Total Probably Rule Mulliiplication Rules — PCANB)= PCBIA) P(A) = PCAIB) P(8) Total Probability Rule? Few any ends A and 8, CFer-ton wes) peg)= PCBNA)+ PCBOA!) = PCBIA) PCA)+ P(B/A!) PLA!) (few vuliple weds) TP Ey, Ey). Ey, ven mutually exclusive nol exhaustive sets, Thew PCB) = PCBNE,) + PBN Es)+ ---+ P(BNE,) = P(BIE,).PCE:) + P(B/Es).PCE,) +--+ P (B/E,) PEs) NGS! Th the above tot probability rule, a collection of sets 1, Ex,--5En such that E,UE,U--.UE, = is said-ty be exhaustive, Tndsptndince! In some cases, the condlitioal probability of FBIA) “might qual P(B). Tn this speclaf case, hrauleclge thot the outcome ofthe exfurinnend Is in tuted A clot mot ajfect the probability hod thy outcome is in ent B, These cases Lea! -to Om jmportat definition’ Trdsfindina of tos Utots! Tuo exoasycit ndeptodiat if ong one of the Foflaoing equivalent stottrmends is tAur! DA NIShYy GUPte 19 ch PCA) = Pca) @ (2) P(BiA) = PCB) (3) PCANB)= PCA). PCB) “Nole? Tf A dnd B ane inckependud events, thers PCAN BI )= PLAN)PCB') be, Alamel B! are alse indipieded Prot Wehave poping!) = pCAus)! — (by Demergar's Law) = 1 PAUB) E+ PeaertA =I] = IPO PCB)+PCANB) (by aollitin ne) = I- P(AI-P(B) + P(A). PLB) (Ae ee = 1(l-P¢ay)— P68) (1- POA) indepemslet) = (1-P(A)) (1-P0B)) = PCAY).erg!) Hence Provenl Ardeponstince the events Er, Ey, Ea are independant parol. Cmubkige wets) only if Tor any subse of thie erent Ey Eig 1 Eine , PCEu NE, N---NEy,) = P(Ein) x P(E,)x ~--x PCEY.) Exdinple TF PCAIB)=03, P(B) = 0.8 and P(A) = 0/3, are the Cents Band the complement of A indepenrclet? “° PCAIB)= PCA) = 0:3 Sal +f dnt Bane inclepondest P(A PCB) = PCAN) () Now, P(A'NB)= P(B)-PCANB) = PCB)— PLA): PCB) byt) = PCe)(I-P(ay] = ' i 8) [I-P(A)] = PCB).PrA!) = Alans Bae fh Nisha Gupta a5) eo Note: Inthe above exavmplle, we have also shown thap if Aone & ore indeberdest, tht Alan! B ane also inclebendenct. Sivailandy , we can shots that i# A anol B are inclopenclet, thin A ancl Bi are also Inclepenclest. Qu: The -Follasing rable summarizes the analysis of Somples OF galwanizest steel fr coating. votight Onel surface Rouges! coating weight hi Low surface roughness high 1g 16 Jow 88 34 (a) Tf -th coating. wwught of a Sdmple is high, what isthe probability thod the swface Agughness is Nios ¢ ? CL) Th the surface roughness of 4 sumple is Nob , , lahat Is th “probability thot the coading. weight is Nios ? () Titik surface houghness Of 4 Sdmple Is Leto, what isthe “probability thot the coating we okt is fav) 7 Sol Let A oltmote the evtnt that a Saople | har ne Coad ng totight and fet Bo» a i hot Surpace AOU rasS, We wart -totind PCB/A) , PCA/B) omd PCA‘/B') There can he-firol by sto oliffered ways. From -the qe table , we have Dhv NiShy Gupte Coaking. weight @D hi Low | Toast Stuf QcLh sughrws sly gh dew Total . . ight highs ( PCB/A) = 12 (7! of the-ponts With coating weight " vn i ( cto the 70 .0f high coaetirng Swe = OD Aoughnass is 12) P(B/A) = PCANB) = T50 1a 2 ge PCA) Cb) PCA/B)= 2 (1 Of the pants with surface Aoughress _ aa higa (a8-poab) the no of he, coating. ow couged is 12) Pcafa)= pcane) . pes PC8) ae G7 'So PCA e)= 34 ( Of He forbs with sex surface, [22 Ss cughnass low (129 ants) therne. (c) = IL él of tow coating. diag is34) or PcA'yg!) = ing) _ St 7 CAB) = PCA'NB') = TE _ 3427 Pce') 122 a2 GI i Dh. Nishu Gupta (2) ‘Qust (Machining Stoyt) The probability thatthe First stage Gf 0 rumerically conballicl machining. ofptration For Nigh pn pistons mets specifications is 0-40. Foudures are due to 7 metal Voniations, fixture ab gnc, Cutting. bloole conelttins. Giiven-thot the First stage wments spectFications , “the probally thot a second stage of machining mets specications is a buhay is-the ‘probability that beth stage, meat speeFeatios 7 So Let AQnol B cluwote the events hod the First nol seconal, stoges meet spectfications, Aaspectively . ". PCANB) = PCBIA): PAD (by multiplication Atte = 095(0-40) = 0855 Bue (Semiconduckr Failury) Codinuing with semicndutter rmomucfacunng assume thefollowing -probabiliti tor “product foilure subjectto fovels of contomnation iw mamufacwung | Protabibiyty foiluee Lewel of Qrnlomination 0-lo , High 0-01 Meclitns 0-001 Loto Ina particu production rim, oie of the chips ane subjected to high Levebop Contamination, Soi, to meclivn fovels of Contamination, anol So%/e to tow lovdls oF Covdaminaton. Whot Ls the Probability that 4 product Uning ome of these chips Fails ? (23) / Sol Ler F dimote the eveet thatthe Brocluey-foils amd Lot H dumobethe evwd that a chip is exposed te Nok deel of otennele Lowa » : M » ” » yD ” ae Lo a ee, Don Then, PCE) = PCF[HI.PCH) + PCF/M). PCM) PCF/L)-P(L] —(1) (by Total Probability. Atle For teecllrple wets ) Naw, accardirg to given olata Level of Contamination Probability of Failure Probability. af Level Hi la = 5 =P(H igh 0:0 = PCF/H) “i= 020 pH) Meditin 0-0) = PCF/M) Be = 0:30= PUM) Lows g-001 = PCF/L) se = 0-50= PCL) lwo J From equim (0), PCF) =(0+10) (0-20) + (0-01) (0-30) + (0-001) (0-50) = 00935 A Que (Poralled Crcuit) The following cincuit operates oly if there Is a Path of Functional ie Form Left to Aight. The probability that each dewice functions is shown onthe qrabh | Assume thot clwies fail independidly . Who is Ge Probability thatthe GAuuit operates? BEI L—— L a—4 -—— 0.45 Dh Nishy Gupters yy Sof / Sh Ltt T deotethe tnt that the tof olovice. operate ama) RB» ” » » bottom » Es Thee is a path if ad feast one chovice operates. The probability. that the Circuits oprate = P(TUB) = 1 pC Tue)! “(7 PLE) = ce’) = 1—pcr'ne') ( by DeMorgan's 4m) = Ie PCr.Pea!) (7 Tans! 8 axe inclapewelted =) Tlanol BI ae asso indlape-dh-t) = |-(l-0-95) (1-099) = |oos*=oaq7s A Qu A batch of 500 Containers For Frozen onsnge jules coxtalns —five that one olefective. Two ane selected, a pandlom, withad ALplacetnt From the batch. Let A dno B denote tho tutnts hod the first anol Seconol containers select col anc olefechve, pespecively . (a) Ape Aanst B indepenclit events ( (b) EfthL sampling. wore done with ptplacement, would, A anol B be inclepenolent ? Also, Final the Probability that both pants ane clapective. in 2ach case. _Sof @ Tn case of without frplacement the events A anol B are not indlependet Ano in-this case PCANB) = PCA)-P(B/A) = 5 x4 = 4 = 00° 499 «49900-12475 / Dh Nishy Gupte co 5) [ (0) Tn cane of taith Aeplaciment the Quink A dno! Bane / indepencbed Arnel in His PCANB)= PCA). PCB) (0 PCB/A)= P¢8)) = ats eos +0001 Bus A problem in mathimodics Is given to two stuolents A dns B.ThuA Aespective Probability of soluing it ane td LT both of thom try. to solve. the problem rele- pendently , Fira the probability that (a) Exactly one of thine solve the problen Ch) Nene of them solve the problow (c) Problem is solveot Sol Let us Aspire the Foblorst tuuss | Ey! The Probltm is solvect by studi A Eg! The » noo» » 2 8 Then PCE = PCEa= = Mi BGEW alae aes PCE) = IHL ae 2 27 3 OF Ca) PC Exactly one of thin solve the proble,.) PCE: Es, )+ PCEWN Ex) = PCE,). PCE, ) + PCE). P(Ea) rdtgg-d (4) PO Nome of thten solve. it) crak PCE,NE,) = PCE): PCE) CC) PC Problem is solved) = I-PCEMEs) Dh Nishu Guble (26) Note: Fyemthe mulfrfLeafion ace, PCANB) = PCA/8).P(B) ——() = P(B/Alp(A) ———™ PcaB) = (An B) byw) gts) This result is usefiet to solwe For p(A/B) SP Pcarp)= P(B/A). PCA) for PtB) > 0 PCB) Bay's Theowen: Hf Ey, Ea)---» Epa k routually exclusive ond exhaustive events anol A is dup wet, PCE IA) = PALE)» PCE) for P(A) 70 5 PCAIE:) PCEL) anal jehorck l= rte dimaminater ,B PCALED): PCEL)= PCA) interns off P(B/A). Note: Que Sufpose phat PCA/B) = 0°7) P(A)=0°5 anol P(B)= 0-2. Deferrine P(8/A). Sol P(B/A) = PCA/B). PCB) : P(A) PBA) = g7x0.0 2 Ol 2028 sO i Our A Pinter manufactur obtainssl the Po llawing probabilitie, From a ladabase of-test Arsults. Prindtr Failures ane associoted Wtth thre types of problems: harduare , software, dnol other (such as conmecdors) With probabilities O-1, 0.6 anal 0-3 Mepectivaly The probability of & Panter failure divin a harcore Probl. Dai Nishu Gupte (7) is 0.9, given a softourt problin is 0.2 drt Yvten anny oth type of problien Is 0-5. TF a customer entias the mauner fates website to ciagmose aprintr failure, what is the most Likely cause of the problem ( Sol Leap the events Hy S anal 0 cltnate q hardiare , softer or other problem prospectively anal Let F clinote a Printer failure The most Likely cause of the problin fs the one “thad comasponds to -the Longest of PCH/F), P( S/F) eee PCO/F). Here given PCH)sot, P(S)=0-6 , PCO)=0S PCE/M)= 0-9 , PCE/s)20-2, PCF/O)= 05 Naw, pce = PCE/H).PcH)-+ PCFISPCS) + PC F/O). PCO) = (0-9) (0-1) + (0-2) (0-6) + (05) (0.3) = 036 By Bays's theorem, PCH/p) = PCEIH).PlH) = (O-V0-) = 0-250 PCF) 0-36 PC sre) = PCRIS)PCS) - (0:2) (06) = 0-333 PCF) 0-36 PCO/F) = Pcr/O)-P(O) = (4-S)fo3) = 947 PCF) Ce “2 PCOPE) is Dangest, the most Likely couse of the Probl iS inthe other categarg.. Note! The above ‘Problam is an example of So re woth one used onthe web sites of high techn alogep minitacts allow customens to qichDy. cliogmase Pavbloms loith produc.

You might also like