Beta Bessel distributions英文文献资料.docVIP

BETABESSELDISTRIBUTIONS ARJUNK.GUPTAANDSARALEESNADARAJAH Received5December2005;Revised4May2006;Accepted7May2006 Threenewdistributions ontheunitinterval [0,1]areintroduced which generalize the standardbetadistribution. Thesedistributions involvetheBesselfunction.Expressionis derivedfortheirshapes,particularcases,andthenthmoments.Estimationbythemethod of maximum likelihood and Bayes estimation are discussed. Finally, an application to consumer priceindicesisillustrated toshowthattheproposed distributions arebetter modelstoeconomicdatathanonebasedonthestandardbetadistribution. Copyright?2006HindawiPublishingCorporation.Allrightsreserved. 1.Introduction Betadistributionsareveryversatileandavarietyofuncertaintiescanbeusefullymodeled by them. Many of the ?nite range distributions encountered in practice can be easily transformed into the standard distribution. In reliability and life testing experiments, manytimesthedataaremodeledby?niterangedistributions, see,forexample,[2]. ArandomvariableXissaidtohavethestandardbetadistribution withparametersν andμifitsprobabilitydensityfunction(pdf)is xν ?1(1?x)μ?1 f(x)= for0x1,ν0,andμ0,where (1.1) B(ν,μ) B(a,b)= ta?1(1?t)b?1 1 dt (1.2) 0 denotesthebetafunction.Manygeneralizationsof(1.1)involvingalgebraic,exponential, and hypergeometric functions havebeen proposed intheliterature. Some ofthese are (see[6,Chapter25]and[5]forcomprehensiveaccounts) HindawiPublishingCorporation InternationalJournalofMathematicsandMathematicalSciences Volume2006,ArticleID16156,Pages1–14 DOI10.1155/IJMMS/2006/16156  2 BetaBesseldistributions (i)thefour-parametergeneralizationgivenby  a?1 1?d?c x?c b?1 1 x?c f(x)=(d?c)B(a,b) d?c (1.3) forc≤x≤d,a0,andb0(see[7,Section 3.2]forareparameterization of this); (ii)theMcDonaldandRichards[9,13]betadistributiongivenby  b?1 pxap ?1 1?( x/q)p qapB(a,b) f(x)= (1.4) (1.5) for0≤x≤q,a0,b0,p0,andq0; (iii)theLibbyandNovick[8]betadistributiongivenby λaxa ?1(1?x)b?1 f(x)=B(a,b) 1?(1?λ)x