- Information
- AI Chat
Prob3160ch4 - math
Differential Equation l (MATH 2230)
Central Luzon State University
Preview text
CHAPTER 4
Conditionalprobability
- Denition,Bayes'Ruleandexamples
Supp osethere are 200 men,of which 100 aresmokers,and 100 women,of which 20 are smokers. Whatistheprobabilitythatap ersonchosenatrandomwillb easmoker? The answeris 120 / 300. Now,letusask,whatistheprobabilitythatap ersonchosenatrandom isasmokergiventhatthep ersonisawomen? Onewouldexp ecttheanswertob e 20 / 100 anditis.
Whatwehavecomputedis
numb erofwomensmokers numb erofwomen
=
numb erofwomensmokers/ 300 numb erofwomen/ 300
,
whichisthesame asthe probabilitythat ap ersonchosenat randomisawomananda smokerdividedbytheprobabilitythatap ersonchosenatrandomisawoman.
Withthisinmind,wegivethefollowingdenition.
Denition4(Conditionalprobability)
IfP(F)> 0 ,wedenetheprobabilityof EgivenFas
P(E|F) :=
P(E∩F)
P(F)
.
NoteP(E∩F) =P(E|F)P(F).
Example4. Supp oseyourolltwodice. Whatistheprobabilitythesumis8?
Solution:therearevewaysthiscanhapp en{(2,6),(3,5),(4,4),(5,3),(6,2)},sotheprob- abilityis 5 / 36. LetuscallthiseventA. Whatistheprobabilitythatthesumis 8 given thattherstdieshowsa3? LetBb etheeventthattherstdieshowsa3(A∩B) istheprobabilitythattherstdieshowsa 3 andthesumis8,or 1 / 36. P(B) = 1/ 6 , so
P(A|B) = 11 // 366 = 1/ 6.
Example4. Supp oseab oxhas 3 redmarblesand 2 blackones. Weselect 2 marbles. Whatistheprobabilitythatsecondmarbleisredgiventhattherstoneisred?
41
42 4
Solution: LetAb etheeventthesecondmarbleisred,andBtheeventthattherstone isred. P(B) = 3/ 5 ,whileP(A∩B)istheprobabilityb otharered,oristheprobability thatwe chose 2 redoutof 3 and 0 blackoutof2. ThenP(A∩B) =
( 3
2
)( 2
0
)
/
( 5
2
)
, andso
P(A|B) = 33 // 105 = 1/ 2.
Example4. Afamilyhas 2 children. Giventhatoneofthechildrenisab oy,whatis theprobabilitythattheotherchildisalsoab oy?
Solution:LetBb etheeventthatonechildisab oy,andAtheeventthatb othchildrenare b oys. Thep ossibilitiesarebb, bg, gb, gg,eachwithprobability 1 / 4 .P(A∩B) =P(bb) = 1/ 4
andP(B) =P(bb, bg, gb) = 3/ 4 .Sotheansweris 13 // 44 = 1/ 3.
Example4. Supp osethetestforHIVis99%accurateinb othdirectionsand0%ofthe p opulationisHIVp ositive. Ifsomeonetestsp ositive,whatistheprobabilitytheyactually areHIVp ositive?
Solution:LetDistheeventthatap ersonisHIVp ositive,andTistheeventthatthep erson testsp ositive.
P(D|T) =
P(D∩T)
P(T)
=
(0)(0)
(0)(0) + (0)(0)
≈23%.
Ashortreasonwhythissurprisingresultholdsisthattheerrorinthetestismuchgreater thanthep ercentageofp eoplewithHIV osethatwehave 1000 p eople. Onaverage, 3 ofthemwillb eHIVp ositiveand 10 willtestp ositive. Sothe chancesthatsomeonehasHIVgiventhatthep ersontestsp ositiveisapproximately 3 / 10. Thereasonthatitisnotexactly 0. 3 isthatthereissomechancesomeonewhoisp ositive willtestnegative.
Supp oseyouknowP(E|F)andyouwanttondP(F|E).Recallthat
P(E∩F) =P(E|F)P(F),
andso
P(F|E) =
P(F∩E)
P(E)
=
P(E|F)P(F)
P(E)
Example4. Supp ose36%offamiliesownadog, 30%offamiliesownacat,and22% ofthefamiliesthathaveadogalsohaveacat haveacat?
Solution:LetDb ethefamiliesthatownadog,andCthefamiliesthatownacat givenP(D) = 0. 36 ,P(C) = 0. 30 ,P(C |D) = 0. 22 WewanttoknowP(D|C). Weknow
44 4
Hereisanotherexamplerelatedtoconditionalprobability,althoughthisisnotanexample ofBayes'rule. ThisisknownastheMontyHal lproblemafterthehostoftheTVshowin the60scalledLet'sMakeaDeal.
Example 4. Therearethreedo ors, b ehindonea nicecar,b ehindeach of theother twoagoateatingabaleofstraw. Youcho oseado or. ThenMontyHallop ensoneofthe otherdo ors,whichshowsabaleofstraw. Hegivesyoutheopp ortunityofswitchingtothe remainingdo or?
Solution: Let'ssupp oseyoucho osedo or1,sincethesameanalysisapplieswhicheverdo or youchose. Strategyoneistostickwithdo or1. With probability 1 / 3 youchosethecar. MontyHallshowsyouoneoftheotherdo ors,butthatdo esn'tchangeyourprobabilityof winning.
Strategy 2 istochange. Let'ssaythecarisb ehinddo or1,whichhapp enswithprobability 1/3 ors,saydo or2 eagoat,soyou switchtodo or3,andlose. Thesameargumentappliesifheshowsyoudo or3. Supp ose thecarisb ehinddo or2. Hewillshowyoudo or3,sincehedo esn'twanttogiveawaythe car or 2 andwin. Thishapp enswithprobability 1 / 3. Thesameargument appliesifthecarisb ehinddo or3/3andlosewithprobability 1/3 2 ismuchsup erior.
Aproblemthatcomesupinactuarialsciencefrequentlyisgambler'sruin.
Example4 (Gambler'sruin). Supp oseyouplaythegamebytossingafaircoinre- p eatedlyandindep endently,youwinadollar,andifitcomesuptails, youloseadollar. Supp oseyoustartwith$50'stheprobabilityyouwillgetto$ withoutrstgettingruined(runningoutofmoney)?
Solution:itiseasiertosolveaslightlyharderproblem edescrib edashaving probability 1 / 2 ofwinning 1 dollarandaprobability 1 / 2 oflosing 1 dollar egins withagivennumb erof dollars,andintendstoplay thegamerep eatedly untiltheplayer eithergo esbrokeorincreaseshisholdingstoNdollars.
Foranygivenamountnofcurrentholdings,theconditionalprobabilityofreachingNdollars b eforegoing brokeisindep endent ofhow weacquiredthendollars, sothereisaunique probabilityP(N|n)ofreachingN ontheconditionthatwecurrentlyholdndollars. Of course,foranyniteNweseethatP(N|n) = 0andP(N|N) = 1. Theproblemisto determinethevaluesofP(N|n)fornb etween 0 andN.
WeareconsideringthissettingforN= 200,andwewouldliketondP(200|50). Denote y(n) :=P(200|n), whichistheprobabilityyougetto 200 withoutrstgettingruinedif youstartwithndollars. Wesawthaty(0) = 0andy(200) = 1. Supp osetheplayerhas ndollarsatthemoment,the nextroundwill leavetheplayerwitheithern+ 1orn− 1 dollars,b othwithprobability 1 / 2 .Thusthecurrentprobabilityofwinningisthesameasa weightedaverageoftheprobabilitiesofwinninginplayer'stwop ossiblenext states
4.1,BAYES'RULEANDEXAMPLES 45
have y(n) = 12 y(n+ 1) + 12 y(n−1).
Multiplyingby2,andsubtractingy(n) +y(n−1)fromeachside,wehave
y(n+ 1)−y(n) =y(n)−y(n−1).
Thissaysthatslop esofthegraphofy(n)ontheadjacentintervalsareconstant(rememb er thatxmustb eaninteger).Inotherwords,thegraphofy(n)mustb ealine(0) = 0 andy(200) = 1,wehavey(n) =n/ 200 ,andthereforey(50) = 1/ 4.
Anotherwaytoseewhatthefunctiony(n)istousethetelescopingsumasfollows
(4.1) y(n) =y(n)−y(0) = (y(n)−y(n−1)) +...+ (y(1)−y(0))
=n(y(1)−y(0)) =ny(1).
sincetheallthesedierencesarethesame,andy(0) = 0. Tondy(1)wecanusethefact thaty(200) = 1,soy(1) = 1/ 200 ,andthereforey(n) =n/ 200 andy(50) = 1/ 4.
Example4. Supp oseweareinthesamesituation,butyouareallowedtogoarbitrarily farindebt. Letz(n)b etheprobabilityyouevergetto$200ifyoustartwithndollars. Whatisaformulaforz(n)?
Solution:Justasab ove,weseethatzsatisestherecursiveequation
z(n) = 12 z(n+ 1) + 12 z(n−1).
Whatweneedtodeterminenowareb oundaryconditions. Nowthatthe gamblercango todebt,thecondition thatifwestartwith 0 wenevergetto$200,thatis,probabilityof getting$200is 0 ,isnottrue.FollowingEquation4.1(0) 6 = 0weseethat
z(n)−z(0) = (z(n)−z(n−1)) +...+ (z(1)−z(0)) =n(z(1)−z(0)),
therefore z(n) =n(z(1)−z(0)) +z(0).
Ifwedenotea:=z(1)−z(0)andb:=z(0)weseethatasafunctionofnwehave
z(n) =an+b.
Wewouldliketondaandbnow,soforanynwe have 06 z(n) 61 .Thisisp ossibleonlyifa= 0,thatis,
z(1) =z(0),
so
z(n) =z(0)
foranyn(200) = 1,therefore
4.2 47
- Furtherexamplesandapplications
4.2. Examples,basicprop erties,multiplicationrule,lawoftotalprobability.
Example 4. Landonis 80% surehe forgot his textb o okeither at the Unionor in Monteith%surethattheb o okisattheunion,and40%surethatitisinMonteith. GiventhatLandonalreadywenttoMonteithandnoticedhistextb o okisnotthere,whatis theprobabilitythatitisattheUnion?
Solution:CallingU=textbookisattheUnion,andU=textbookisinMonteith,noticethat U⊆McandhenceU∩Mc=U. Thus,
P(U|Mc) =
P(U∩Mc) P(Mc)
=
P(U)
1 −P(M)
=
4 / 10
6 / 10
=
2
3
.
Example 4. SarahandBobdraw 13 cardseachfromastandarddeckof 52. Given thatSarahhasexactlytwoaces,whatistheprobabilitythatBobhasexactlyoneace?
Solution: LetA=Sarahhastwoaces,andletB=Bobhasexactlyoneace. Inorderto computeP(B|A),weneedtocalculateP(A)andP(A∩B). Ontheonehand,Sarahcould haveanyof
( 52
13
)
p ossiblehands,
( 4
2
)
·
( 48
11
)
willhaveexactlytwoacessothat
P(A) =
( 4
2
)
·
( 48
11
)
( 52
13
).
Ontheotherhand,thenumb erofwaysinwhichSarahcanpickahandandBobanother (dierent)is
( 52
13
)
·
( 39
13
)
.Thethenumb erofwaysinwhichAandBcansimultaneouslyo ccur is
( 4
2
)
·
( 48
11
)
·
( 2
1
)
·
( 37
12
)
andhence
P(A∩B) =
( 4
2
)
·
( 48
11
)
·
( 2
1
)
·
( 37
12
)
( 52
13
)
·
( 39
13
).
Applyingthedenitionofconditionalprobabilitywenallyget
P(B|A) =
P(A∩B)
P(A)
=
( 4
2
)
·
( 48
11
)
·
( 2
1
)
·
( 37
12
)/( 52
13
)
·
( 39
13
)
( 4
2
)
·
( 48
11
)/( 52
13
) =
2 ·
( 37
12
)
( 39
13
)
Example 4. Atotalof 500 marriedcouplesarep olledab outtheirsalarieswiththe followingresults
husbandmakeslessthan$25K husbandmakesmorethan$25K wifemakeslessthan$25K 212 198 wifemakesmorethan$25K 36 54
(a)Findtheprobabilitythatahusbandearnslessthan$25K. Solution:
P(husbandmakes <$25K) =
212
500
+
36
500
=
248
500
= 0. 496.
©Copyright 2017 PhanuelMariano,PatriciaAlonsoRuiz,Copyright 2020 MashaGordina
48 4
(b)Findtheprobabilitythatawifeearnsmorethan$25K,giventhatthehusbandearns asthatmuchaswell. Solution:
P(wifemakes >$25K|husbandmakes >$25K) =
54 / 500
(198 + 54)/ 500
=
54
252
= 0. 214
(c)Findtheprobabilitythatawifeearnsmorethan$25K,giventhatthehusbandmakes lessthan$25K. Solution:
P(wifemakes >$25K|husbandmakes <$25K) =
36 / 500
248 / 500
= 0. 145.
Fromthedenitionofconditionalprobabilitywecandeducesomeusefulrelations.
Prop osition4.
LetE, F∈ Fb eeventswithP(E),P(F)> 0 .Then (i) P(E∩F) =P(E)P(F|E), (ii)P(E) =P(E|F)P(F) +P(E|Fc)P(Fc), (iii)P(Ec|F) = 1−P(E|F).
Proof(i)whichisarewritingofthedenitionofconditionalprobability
P(F|E) =P(PE(∩EF)). Letusprove(ii): wecanwriteEastheunionofthepairwisedisjoint
setsE∩FandE∩Fc. Using(i)wehave
P(E) =P(E∩F) +P(E∩Fc) =P(E|F)P(F) +P(E|Fc)P(Fc).
Finally,writingF=EinthepreviousequationandsinceP(E|Ec) = 0,weobtain(iii).
Example4. PhanwantstotakeeitheraBiologycourseoraChemistrycourse. His adviserestimatesthattheprobabilityofscoringanAinBiologyis 45 ,whiletheprobability ofscoringanAinChemistryis 17. IfPhandecidesrandomly,byacointoss,whichcourse totake,whatishisprobabilityofscoringanAinChemistry?
Solution: denotebyBtheeventthatPhan takesBiology,andbyC theeventthat Phan takesChemistry,andbyA=theeventthatthescoreisanA,sinceP(B) =P(C) = 12 wehave
P(A∩C) =P(C)P(A|C) =
1
2
·
1
7
=
1
14
.
TheidentityP(E∩F) =P(E)P(F|E)fromProp osition4(i)canb egeneralizedtoany numb erofeventsinwhatissometimescalledthemultiplicationrule.
50 4
Solution:
(a)denotebyA 1 =theeventthatapolicyholderwil lhaveanaccidentwithinayear,and denotebyA =the eventthata policyholder is accidentprone. ApplyingProp osi- tion4(ii)wehave P(A 1 ) =P(A 1 |A)P(A) +P(A 1 |Ac) (1−P(A)) = 0. 4 · 0 .3 + 0(1− 0 .3) = 0. 26 (b)UseBayes'formulatoseethat
P(A|A 1 ) =
P(A∩A 1 )
P(A 1 )
=
P(A)P(A 1 |A)
0. 26
=
0. 3 · 0. 4
0. 26
=
6
14
.
UsingthelawoftotalprobabilityfromProp osition4'rule,which app earedinProp osition4.
Prop osition4(GeneralizedBayes'rule)
LetF 1 ,... , Fn⊆ S b emutuallyexclusiveandexhaustiveevents,i. S =
⋃n i=1Fi. Then,foranyeventE⊆Sandanyj= 1,... , nitholdsthat
P(Fj|E) =
P(E|Fj)P(Fj) ∑n i=1P(E|Fi)P(Fi)
Example4. Supp oseafactoryhasmachinesI,I I,andI I Ithatpro duceiSung phones. Thefactory'srecordshowsthat
MachinesI,I IandI I Ipro duce,resp ectively,2%,1%,and3%defectiveiSungs. Outofthetotalpro duction,machinesI,I I,andI I Ipro duce,resp ectively,35%,25%and 40%ofalliSungs.
AniSung isselectedatrandomfromthefactory.
(a)WhatisprobabilitythattheiSungselectedisdefective? Solution:Bythelawoftotalprobability, P(D) = P(I)P(D|I) +P(II)P(D|II) +P(III)P(D|III)
= 0. 35 · 0 .02 + 0. 25 · 0 .01 + 0. 4 · 0 .03 =
215
10 , 000
.
(b)GiventhattheiSungisdefective,whatistheconditionalprobabilitythatitwaspro- ducedbymachineI I I? Solution:ApplyingBayes'rule,
P(III|D) =
P(III)P(D|III)
P(D)
=
0. 4 · 0. 03
215 / 10 , 000
=
120
215
.
Example4. Inamultiplechoicetest,astudenteitherknowstheanswertoaquestion orshe/hewillrandomlyguessit. Ifeachquestionhasmp ossibleanswersandthestudent
4.2 51
knowstheanswertoaquestionwithprobabilityp,whatistheprobabilitythatthestudent actuallyknowstheanswertoaquestion,giventhathe/sheanswerscorrectly?
Solution:denotebyKtheeventthatastudentknowstheanswer,andbyC theeventthat astudentanswerscorrectly. ApplyingBayes'rulewehave
P(K|C) =
P(C|K)P(K)
P(C|K)P(K) +P(C|Kc)P(Kc)
=
1 ·p 1 ·p+m 1 (1−p)
=
mp 1 + (m−1)p
.
4.3 53
Exercise4. Amultiplechoiceexamhas 4 choicesforeachquestion. Thestudenthas studiedenoughsothattheprobabilitytheywillknowtheanswertoaquestionis0,the probabilitythat the studentwill b eable toeliminate onechoiceis0, otherwise all 4 choicesseemequallyplausible. Iftheyknowtheanswertheywillgetthequestioncorrect. Ifnottheyhavetoguessfromthe 3 or 4 choices measurewhatthestudentknows,andnothowwelltheycanguess. Ifthestudentanswers aquestioncorrectlywhatistheprobabilitythattheyactuallyknowtheanswer?
Exercise4. Ablo o dtestindicatesthepresenceofAmyotrophiclateralsclerosis(ALS) 95%ofthetimewhenALSisactuallypresent 0%ofthetimewhenthediseaseisnotactuallypresent. Onep ercentofthep opulation actuallyhasALS ersonactuallyhasALSgiventhatthe testindicatesthepresenceofALS.
Exercise 4. Asurveyconductedinacollegefoundthat 40%of thestudentswatch showAand17%ofthestudentswhofollowshowA,alsowatchshowB,20% ofthestudentswatchshowB.
(1)Whatistheprobabilitythatarandomlychosenstudentfollowsb othshows? (2)Whatis theconditional probabilitythat the studentfollowsshow Agiven that she/hefollowsshowB?
Exercise4. UseBayes'formulatosolvethefollowingproblem orthasproblems withbirds. Iftheweatherissunny, theprobabilitythattherearebirdson therunwayis 1 / 2 ;ifitiscloudy,butdry,theprobabilityis 1 / 3 ;andifitisraining,thentheprobability is 1 / 4. Theprobabilityofeachtyp eoftheweatheris 1 / 3. Giventhatthebirdsareonthe runaway,whatistheprobability
(1)thattheweatherissunny? (2)thattheweatheriscloudy(dryorrainy)?
Exercise4. Supp oseyoutossafaircoinrep eatedlyandindep endently. Ifitcomesup heads,youwinadollar,andifitcomesuptails,youloseadollar. Supp oseyoustartwith $20$150b eforeyougobroke? (SeeExample4 asolution).
Exercise∗4. Supp oseweplaygambler'sruingameinExample4, butratherinsuchawaythatyouwinadollarwithprobabilityp,andyouloseadollarwith probability 1 −p, 0 < p < 1. FindtheprobabilityofreachingN dollarsb eforegoingbroke ifwestartwithndollars.
Exercise∗ 4. Supp oseF isanevent, anddenePF(E) :=P(E|F). Showthat the conditionalprobabilityPF isaprobabilityfunction,thatis,itsatisestheaxiomsofproba- bility.
54 4
Exercise∗ 4. ShowdirectlythatProp osition2 conditionalprobability PF. Inparticular,foranyeventsEandF
E(Ec|F) = 1−E(E|F).
56 4
SolutiontoExercise4: denotebyM theeventanAmericanisaman,byCtheevent anAmericaniscolorblind
P(M|C) =
P(C|M)P(M)
P(C|M)P(M) +P(C|Mc)P(Mc)
=
(0) (0)
(0) (0) + (0) (0)
≈ 0. 9505.
SolutiontoExercise4(A):letHb etheeventabulbworksover 5000 hours,Xb ethe eventthatabulbcomesfromfactoryX,andY b etheeventabulbcomesfromfactoryY. Thenbythelawoftotalprobability
P(H) =P(H|X)P(X) +P(H|Y)P(Y) = (0) (0) + (0) (0) = 0. 974.
SolutiontoExercise4(B):ByPart(a)wehave
P(Y |H) =
P(H|Y)P(Y)
P(H)
=
(0) (0)
0. 974
≈ 0. 39.
SolutiontoExercise4(C):WeagainusetheresultfromPart(a)
P(X|Hc) =
P(Hc|X)P(X) P(Hc)
=
P(Hc|X)P(X) 1 −P(H)
=
(1− 0 .99) (0)
1 − 0. 974
=
(0) (0)
0. 026
≈ 0. 23
SolutiontoExercise4:denotebyDtheeventthataboltisdefective,Atheeventthat aboltisfrommachineA,byBtheeventthataboltisfrommachineC. ThenbyBayes' theorem
P(A|D) =
P(D|A)P(A)
P(D|A)P(A) +P(D|B)P(B) +P(D|C)P(C)
=
(0) (0)
(0) (0) + (0) (0) + (0) (0)
= 0. 362.
SolutiontoExercise4: LetC b etheeventastudentgivesthecorrectanswer,Kb e theeventastudentknows thecorrectanswer,Eb etheeventastudentcaneliminateone incorrect answer, andG b e theevent astudent havetoguessananswer. UsingBayes'
4.4 57
theoremwehave
P(K|C) =
P(C|K)P(K)
P(C)
=
P(C|K)P(K)
P(C|K)P(K) +P(C|E)P(E) +P(C|G)P(G)
=
1 · 12
1 · 12 + 13 · 14 + 14 · 14
=
24
31
≈. 774 ,
thatis,approximately 77 .4%ofthestudentsknowtheansweriftheygivethecorrectanswer.
SolutiontoExercise4:Let+denotetheeventthatatestresultispositive,andbyD theeventthatthediseaseispresent. Then
P(D|+) =
P(+|D)P(D)
P(+|D)P(D) +P(+|Dc)P(Dc)
=
(0) (0)
(0) (0) + (0) (0)
= 0. 657.
SolutiontoExercise4:itisclearthatPF(E) =P(E|F)isb etween 0 and 1 sincethe right-handsideoftheidentitydeningPF is. Toseethesecondaxiom,observethat
PF(S) =P(S|F) =
P(S∩F)
P(F)
=
P(F)
P(F)
= 1.
Nowtake{Ei}∞i=1, Ei∈ Ftob epairwisedisjoint,then
PF
(∞
⋃
i=
Ei
)
=P
(∞
⋃
i=
Ei|F
)
=
P((
⋃∞
i=1Ei)∩F) P(F)
=
P(
⋃∞
i=1(Ei∩F)) P(F)
=
∑∞
i=1P(Ei∩F) P(F)
=
∑∞
i=
P(Ei∩F) P(F)
=
∑∞
i=
PF(Ei).
Inthisweusedthedistributionlawforsets(E∪F)∩G= (E∩G)∪(F∩G)andthefact that{Ei∩F}∞i=1arepairwisedisjointaswell.
Prob3160ch4 - math
Course: Differential Equation l (MATH 2230)
University: Central Luzon State University
- Discover more from: