Scribd
Upload
67 views8 pages

Understanding Beta Distribution

The Beta distribution is a continuous probability distribution used to model uncertainty about the probability of success of an experiment. It has two parameters and is defined on the interval [0,1]. The distribution takes on different forms depending on the number of successes and failures observed in independent trials of a probabilistic experiment where the probability of success is unknown. Specifically, observing x successes out of n trials results in a Beta distribution with parameters x and n-x.

Uploaded by

nimesh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
67 views8 pages

Understanding Beta Distribution

The Beta distribution is a continuous probability distribution used to model uncertainty about the probability of success of an experiment. It has two parameters and is defined on the interval [0,1]. The distribution takes on different forms depending on the number of successes and failures observed in independent trials of a probabilistic experiment where the probability of success is unknown. Specifically, observing x successes out of n trials results in a Beta distribution with parameters x and n-x.

Uploaded by

nimesh
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
  • Beta Distribution Overview
  • Expected Value and Variance
  • Higher Moments
  • Moment Generating Function
  • Characteristic and Distribution Functions
  • Relations to Other Distributions
  • Solved Exercises

Statlect TheDigitalTextbook

Go

Search

Index >Probabilitydistributions

Betadistribution
[Link]
modelone'suncertaintyabouttheprobabilityofsuccessofanexperiment.
Supposeaprobabilisticexperimentcanhaveonlytwooutcomes,eithersuccess,withprobability ,orfailure,with
probability

.Supposealsothat [Link]

canbedescribedbyassigningto auniformdistributionontheinterval

.Thisisappropriatebecause ,beinga

probability,cantakeonlyvaluesbetween and furthermore,theuniformdistributionassignsequalprobability


densitytoallpointsintheinterval,whichreflectsthefactthatnopossiblevalueof is,apriori,deemedmorelikelythan
[Link],supposethatweperform independentrepetitionsoftheexperimentandweobserve successes
and

[Link],wenaturallywanttoknowhowweshouldrevisethedistribution

initiallyassignedto ,[Link]
otherwords,wewanttocalculatetheconditionaldistributionof ,conditionalonthenumberofsuccessesandfailures
[Link],theconditionaldistributionof ,
conditionalonhavingobserved successesoutof trials,isaBetadistributionwithparameters

and

Definition
TheBetadistributionischaracterizedasfollows.
Definition Let [Link]:

Let
functionis

where

.Wesaythat hasaBetadistributionwithshapeparameters and ifitsprobabilitydensity

istheBetafunction.

ArandomvariablehavingaBetadistributionisalsocalledaBetarandomvariable.
Thefollowingisaproofthat

isalegitimateprobabilitydensityfunction.

Proof
Nonnegativitydescendsfromthefactsthat
that

isnonnegativewhen

and

,and

isstrictlypositive(itisaratioofGammafunctions,whicharestrictlypositivewhentheirarguments

arestrictlypositiveseethelectureentitledGammafunction).Thattheintegralof
provedasfollows:

over equals is

wherewehaveusedtheintegralrepresentation

aproofofwhichcanbefoundinthelectureentitledBetafunction.

Expectedvalue
TheexpectedvalueofaBetarandomvariable is

Proof
Itcanbederivedasfollows:

Variance
ThevarianceofaBetarandomvariable is

Proof
Itcanbederivedthankstotheusualvarianceformula(

):

Highermoments
The thmomentofaBetarandomvariable is

Proof
Bythedefinitionofmoment,wehave:

whereinstep

wehaveusedrecursivelythefactthat

Momentgeneratingfunction
ThemomentgeneratingfunctionofaBetarandomvariable isdefinedforany anditis

Proof
Byusingthedefinitionofmomentgeneratingfunction,weobtain

Notethatthemomentgeneratingfunctionexistsandiswelldefinedforany becausetheintegral

isguaranteedtoexistandbefinite,sincetheintegrand

iscontinuousin overtheboundedinterval

Theaboveformulaforthemomentgeneratingfunctionmightseemimpracticaltocompute,becauseitinvolvesan
[Link],thefunction

isafunction,calledConfluenthypergeometricfunctionofthefirstkind,thathasbeenextensivelystudiedinmany
[Link]
softwarepackagesforscientificcomputation.

Characteristicfunction
ThecharacteristicfunctionofaBetarandomvariable is

Proof

Thederivationofthecharacteristicfunctionisalmostidenticaltothederivationofthemomentgeneratingfunction
(justreplace with inthatproof).

Commentsmadeaboutthemomentgeneratingfunction,includingthoseaboutthecomputationoftheConfluent
hypergeometricfunction,applyalsotothecharacteristicfunction,whichisidenticaltothemgfexceptforthefactthat
isreplacedwith .

Distributionfunction
ThedistributionfunctionofaBetarandomvariable is

wherethefunction

iscalledincompleteBetafunctionandisusuallycomputedbymeansofspecializedcomputeralgorithms.
Proof
For

,because cannotbesmallerthan .For

thanorequalto .For

,because isalwayssmaller

Moredetails
InthefollowingsubsectionsyoucanfindmoredetailsabouttheBetadistribution.

Relationtotheuniformdistribution
ThefollowingpropositionstatestherelationbetweentheBetaandtheuniformdistributions.
Proposition ABetadistributionwithparameters

Proof
When

and

,wehavethat

and

isauniformdistributionontheinterval

Therefore,theprobabilitydensityfunctionofaBetadistributionwithparameters
as

and

Butthelatteristheprobabilitydensityfunctionofauniformdistributionontheinterval

canbewritten

Relationtothebinomialdistribution
ThefollowingpropositionstatestherelationbetweentheBetaandthebinomialdistributions.
Proposition Suppose isarandomvariablehavingaBetadistributionwithparameters and .Let be
anotherrandomvariablesuchthatitsdistributionconditionalon isabinomialdistributionwithparameters
and .Then,theconditionaldistributionof given

isaBetadistributionwithparameters

and

Proof
Bycombiningthispropositionandthepreviousone,weobtainthefollowingcorollary.
Proposition Suppose [Link] beanotherrandomvariable
suchthatitsdistributionconditionalon isabinomialdistributionwithparameters and .Then,the
conditionaldistributionof given

isaBetadistributionwithparameters

and

Thispropositionconstitutesaformalstatementofwhatwesaidintheintroductionofthislectureinordertomotivatethe
[Link] independentrepetitionsofarandom
experimenthavingprobabilityofsuccess isabinomialrandomvariablewithparameters and .Accordingtothe
propositionabove,whentheprobabilityofsuccess isaprioriunknownandallpossiblevaluesof aredeemed
equallylikely(theyhaveauniformdistribution),observingtheoutcomeofthe experimentsleadsustorevisethe
distributionassignedto ,andtheresultofthisrevisionisaBetadistribution.

Solvedexercises
Belowyoucanfindsomeexerciseswithexplainedsolutions:
1. Exerciseset1

Thebook
Mostlearningmaterialsfoundonthiswebsitearenowavailableinatraditionaltextbookformat.
Learnmore

Featuredpages

Mainsections

Glossaryentries

Multivariatenormaldistribution

Mathematicaltools

Criticalvalue

Hypothesistesting

Fundamentalsofprobability

Precisionmatrix

Deltamethod

Additionaltopicsinprobability

Meansquarederror

Bernoullidistribution

Probabilitydistributions

Jenseninequality

Bayesrule

Asymptotictheory

TypeIIerror

Independentevents

Fundamentalsofstatistics

Factorial

Explore

About

Share

CentralLimitTheorem

AboutStatlect

Convergenceindistribution

Contacts

Multinomialdistribution

Privacypolicyandtermsofuse

Fdistribution

Websitemap

13
Like

Search   Go Statlect The Digital Textbook Index > Probability distributions Beta distribution The Beta distribution is a cont
where we have used the integral representation a proof of which can be found in the lecture entitled Beta function. Expected 
Higher moments The  ­th moment of a Beta random variable   is Proof By the definition of moment, we have:
where in step   we have used recursively the fact that  . Moment generating function The moment generating function of a Beta
Note that the moment generating function exists and is well defined for any   because the integral is guaranteed to exist and
The derivation of the characteristic function is almost identical to the derivation of the moment generating function (just r
Therefore, the probability density function of a Beta distribution with parameters   and   can be written as But the latter i
Featured pages Multivariate normal distribution Hypothesis testing Delta method Bernoulli distribution Bayes rule Independent

You might also like