应用统计决策课件_第1页
应用统计决策课件_第2页
应用统计决策课件_第3页
应用统计决策课件_第4页
应用统计决策课件_第5页
已阅读5页,还剩21页未读 继续免费阅读

下载本文档

版权说明:本文档由用户提供并上传,收益归属内容提供方,若内容存在侵权,请进行举报或认领

文档简介

应用统计决策StatisticalDecision

FeaturesofDecisionMaking

决策特征ListAlternativeCoursesofAction (PossibleEventsorOutcomes)DeterminePayoffs? (AssociateaPayoffwithEachEventorOutcome)AdoptDecisionCriteria (EvaluateCriteriaforSelectingtheBestCourseof Action)

ListPossibleActionsorEvents

可能行为与事件清单

PayoffTableDecisionTreeTwoMethodsofListingEvent(Ei)CoolWeather(E1) x11=$50

x12=$100WarmWeather(E2) x21=200 x22=125

PayoffTable

盈利表Considerafoodvendordeterminingwhethertosellsoftdrinksorhotdogs.CourseofAction(Aj)SellSoftDrinks(A1)SellHotDogs(A2)

xij

=payoff(profit)foreventiandactionjDecisionTree(决策树):ExampleSoftDrinksFoodVendorProfitTreeDiagramHotDogsCoolWeatherCoolWeatherWarmWeatherWarmWeatherx11=$50

x21=200x22=125x12=100OpportunityLoss(机会损失):Example

Highestpossibleprofitforanevent

Ei -Actualprofitobtainedforanaction

Aj

OpportunityLoss(lij

)Event:CoolWeather

Action:SoftDrinks Profit:$50AlternativeAction:HotDogs

Profit:$100 OpportunityLoss=$100-$50=$50OpportunityLoss:TableEvent OptimalProfitof

SellSoftDrinksSellHotDogs ActionOptimal Action

Cool Hot

100

100-50=50

100-100=0Weather

Dogs

Warm Soft

200

200-200=0

200-125=75WeatherDrinks

AlternativeCourseofActionDecisionCriteria

决策准则ExpectedMonetaryValue(EMV)

Theexpectedprofitfortakinganaction

AjExpectedOpportunityLoss(EOL)

Theexpectedlossfornottakingaction

AjExpectedValueofPerfectInformation(EVPI)

Theexpectedopportunitylossfromthebest decision

DecisionCriteria--EMV

期望货币价值

ExpectedMonetaryValue(EMV)

Sum

(monetarypayoffsofevents)

(probabilitiesoftheevents)Xij

Pi

Vj

NEMVj

=expectedmonetaryvalueofactionjxi,j

=payoffforaction

j

andeventiPi

=probabilityofevent

i

occurringi=1DecisionCriteria--EMVTableExample:FoodVendor

Pi

Event Soft

xijPi

Hot

xijPi

Drinks Dogs.50Cool $50$50

.5=$25$100$100

.50=$50.50Warm $200$200

.5=100$125$25

.50=62.50EMVSoftDrink=$125EMVHotDog=$112.50BetteralternativeDecisionCriteria--EOL

期望机会损失

ExpectedOpportunityLoss(EOL)Sum

(opportunitylossesofevents)

(probabilitiesofevents)

Lj

lijPiEOLj

=expectedmonetaryvalueofactionjli,j

=payoffforactionjandevent

iPi

=probabilityofeventi

occurringi=1NDecisionCriteria--EOLTableExample:FoodVendor

Pi

EventOpLoss

lijPi

OPLoss

lijPi

SoftDrinks HotDogs.50Cool $50 $50

.50=$25$0 $0

.50=$0.50Warm 0 $0

.50=$0$75$75

.50=$37.50

EOLSoftDrinks=$25EOLHotDogs=$37.50BetterChoiceDecisionCriteria--EVPI

完美信息期望价值ExpectedValueofPerfectInformation(EVPI)TheexpectedopportunitylossfromthebestdecisionRepresentsthemaximumamountyouarewilling topaytoobtainperfectinformation

ExpectedProfit

UnderCertainty

-

ExpectedMonetaryValueoftheBestAlternative

EVPI(shouldbeapositivenumber)

EVPIComputationExpectedProfitUnderCertainty

=.50($100)+.50($200) =$150ExpectedMonetaryValueoftheBestAlternative =$125EPVI=$25Themaximumyouwouldbewillingtospendtoobtainperfectinformation.TakingAccountofVariability:FoodVendor

2

forSoftDrink =(50-125)2

.5+(200-125)2

.5=5625

forSoftDrink=75CVforSoftDrinks=(75/125)

100%=60%

2forHotDogs=156.25

forHotdogs=12.5CVforHotdogs=11.11%ReturntoRiskRatio

收益风险比Expressestherelationshipbetweenthereturn(payoff)andtherisk(standarddeviation).RRR=ReturntoRiskRatio=RRRSoft

Drinks=125/75=1.67RRRHotDogs=9YoumightwishtochooseHotDogs.AlthoughSoftDrinkshavethehigherExpectedMonetaryValue,HotDogshaveamuchlargerreturntoriskratioandamuchsmallerCV.DecisionMakingwithSampleInformationPermitsRevisingOldProbabilitiesBasedonNewInformation

NewInformationRevisedProbabilityPriorProbability校正AdditionalInformation:WeatherforecastisCOOL. Whentheweatheriscool,theforecasterwascorrect80%ofthetime.Whenithasbeenwarm,theforecasterwascorrect70%ofthetime.

RevisedProbabilities(校正概率)

Example:FoodVendorPriorProbability先验概率F1=CoolforecastF2=WarmforecastE1=CoolWeather=0.50E2=WarmWeather=0.50P(F1|E1)=0.80 P(F1|E2)=0.30

RevisingProbabilitiesExample:FoodVendorP(E1|F1)

=P(cool)

P(

coolforecast|cool)P(coolforecast)==.73(.50)(.80)(.80)(.50)+(.30)(.50)RevisedProbability

P(F1|E1)=0.80P(F1|E2)=0.30E1=0.50E2=0.50P(E2|F1)

=P(warm)

P(coolforecast|warm)P(coolforecast)=.27RevisedEMVTableExample:FoodVendor

Pi

Event Soft

xijPi

Hot

xijPi

Drinks Dogs.73Cool $50$36.50 $100$73.27Warm $20054 125 33.73EMVSoftDrink=$90.50EMVHotDog=$106.75BetteralternativeRevisedEOLTable Example:FoodVendor

Pi

EventOpLoss

lijPi

OPLoss

lijPi

SoftDrink HotDogs.73Cool $50 $36.50 $0 0.27Warm 0 $0 75 20.25

EOLSoftDrinks=36.50EOLHotDogs=$20.25BetterChoiceRevisedEVPIComputationExpectedProfitUnderCertainty

=.73($100)+.27($200) =$127ExpectedMonetaryValueoftheBestAlternative =$106.75EPVI=$20.25Themaximumyouwouldbewillingtospendtoobtainperfectinformation.TakingAccountofVariability:RevisedComputation

2

forSoftDrinks =(50-90.5)2

.73+(200-90.5)2

.27=4434.75

forSoftDrinks=66.59CVforSoftDrinks=(66.59/90.5)

100%=73.6%

2forHotDogs=123.1875

forHotdogs=11.10CVforHotdogs=(11.10/106.75)

100%=10.4%RevisedReturntoRiskRatioExpressestherelationshipbetweenthereturn(payoff)andtherisk(standarddeviation

温馨提示

  • 1. 本站所有资源如无特殊说明,都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
  • 2. 本站的文档不包含任何第三方提供的附件图纸等,如果需要附件,请联系上传者。文件的所有权益归上传用户所有。
  • 3. 本站RAR压缩包中若带图纸,网页内容里面会有图纸预览,若没有图纸预览就没有图纸。
  • 4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
  • 5. 人人文库网仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对用户上传分享的文档内容本身不做任何修改或编辑,并不能对任何下载内容负责。
  • 6. 下载文件中如有侵权或不适当内容,请与我们联系,我们立即纠正。
  • 7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。

评论

0/150

提交评论