Beta Exponentiated Mukherjii-Islam Distribution: Mathematical Study of Different Properties

Similar documents
Comparing Different Estimators of three Parameters for Transmuted Weibull Distribution

A new Family of Distributions Using the pdf of the. rth Order Statistic from Independent Non- Identically Distributed Random Variables

Random Variables. ECE 313 Probability with Engineering Applications Lecture 8 Professor Ravi K. Iyer University of Illinois

Functions of Random Variables

Estimation of Stress- Strength Reliability model using finite mixture of exponential distributions

STK4011 and STK9011 Autumn 2016

THE ROYAL STATISTICAL SOCIETY GRADUATE DIPLOMA

A NEW MODIFIED GENERALIZED ODD LOG-LOGISTIC DISTRIBUTION WITH THREE PARAMETERS

Dr. Shalabh Department of Mathematics and Statistics Indian Institute of Technology Kanpur

VOL. 3, NO. 11, November 2013 ISSN ARPN Journal of Science and Technology All rights reserved.

Comparison of Dual to Ratio-Cum-Product Estimators of Population Mean

Lecture Notes Types of economic variables

Statistical hybridization of normal and Weibull distributions with its properties and applications

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

The Generalized Inverted Generalized Exponential Distribution with an Application to a Censored Data

STK3100 and STK4100 Autumn 2018

Lecture 3. Sampling, sampling distributions, and parameter estimation

STK3100 and STK4100 Autumn 2017

ρ < 1 be five real numbers. The

UNIVERSITY OF OSLO DEPARTMENT OF ECONOMICS

Lecture 3 Probability review (cont d)

Measures of Entropy based upon Statistical Constants

Module 7. Lecture 7: Statistical parameter estimation

ESTIMATION OF PARAMETERS OF THE MARSHALL-OLKIN EXTENDED LOG-LOGISTIC DISTRIBUTION FROM PROGRESSIVELY CENSORED SAMPLES

Point Estimation: definition of estimators

D KL (P Q) := p i ln p i q i

CHAPTER 3 POSTERIOR DISTRIBUTIONS

Application of Generating Functions to the Theory of Success Runs

Bayes Estimator for Exponential Distribution with Extension of Jeffery Prior Information

Bayesian Inferences for Two Parameter Weibull Distribution Kipkoech W. Cheruiyot 1, Abel Ouko 2, Emily Kirimi 3

Multivariate Transformation of Variables and Maximum Likelihood Estimation

Chapter 5 Properties of a Random Sample

Two Fuzzy Probability Measures

Analysis of Variance with Weibull Data

å 1 13 Practice Final Examination Solutions - = CS109 Dec 5, 2018

Special Instructions / Useful Data

Chapter 4 Multiple Random Variables

Some Statistical Inferences on the Records Weibull Distribution Using Shannon Entropy and Renyi Entropy

BAYESIAN INFERENCES FOR TWO PARAMETER WEIBULL DISTRIBUTION

CHAPTER VI Statistical Analysis of Experimental Data

Comparison of Parameters of Lognormal Distribution Based On the Classical and Posterior Estimates

X ε ) = 0, or equivalently, lim

Lecture Note to Rice Chapter 8

Estimation and Testing in Type-II Generalized Half Logistic Distribution

Bounds on the expected entropy and KL-divergence of sampled multinomial distributions. Brandon C. Roy

Chapter 5 Properties of a Random Sample

Maximum Likelihood Estimation

STRONG CONSISTENCY FOR SIMPLE LINEAR EV MODEL WITH v/ -MIXING

A New Family of Transformations for Lifetime Data

Summary of the lecture in Biostatistics

1 Solution to Problem 6.40

Lecture 4 Sep 9, 2015

IJOART. Copyright 2014 SciResPub.

2SLS Estimates ECON In this case, begin with the assumption that E[ i

ON THE LOGARITHMIC INTEGRAL

best estimate (mean) for X uncertainty or error in the measurement (systematic, random or statistical) best

CHAPTER 6. d. With success = observation greater than 10, x = # of successes = 4, and

Extreme Value Theory: An Introduction

9.1 Introduction to the probit and logit models

ON BIVARIATE GEOMETRIC DISTRIBUTION. K. Jayakumar, D.A. Mundassery 1. INTRODUCTION

Study of Correlation using Bayes Approach under bivariate Distributions

Continuous Distributions

Bayes Interval Estimation for binomial proportion and difference of two binomial proportions with Simulation Study

LECTURE - 4 SIMPLE RANDOM SAMPLING DR. SHALABH DEPARTMENT OF MATHEMATICS AND STATISTICS INDIAN INSTITUTE OF TECHNOLOGY KANPUR

arxiv: v1 [math.st] 24 Oct 2016

A New Measure of Probabilistic Entropy. and its Properties

Lecture 7. Confidence Intervals and Hypothesis Tests in the Simple CLR Model

Estimation of the Loss and Risk Functions of Parameter of Maxwell Distribution

Chain Rules for Entropy

Lecture 02: Bounding tail distributions of a random variable

Chapter 8: Statistical Analysis of Simulated Data

ESS Line Fitting

X X X E[ ] E X E X. is the ()m n where the ( i,)th. j element is the mean of the ( i,)th., then

TESTS BASED ON MAXIMUM LIKELIHOOD

On generalized fuzzy mean code word lengths. Department of Mathematics, Jaypee University of Engineering and Technology, Guna, Madhya Pradesh, India

BAYESIAN ESTIMATOR OF A CHANGE POINT IN THE HAZARD FUNCTION

Channel Models with Memory. Channel Models with Memory. Channel Models with Memory. Channel Models with Memory

f f... f 1 n n (ii) Median : It is the value of the middle-most observation(s).

A Topp-Leone Generator of Exponentiated Power. Lindley Distribution and Its Application

Random Variate Generation ENM 307 SIMULATION. Anadolu Üniversitesi, Endüstri Mühendisliği Bölümü. Yrd. Doç. Dr. Gürkan ÖZTÜRK.

Chapter 14 Logistic Regression Models

Derivation of 3-Point Block Method Formula for Solving First Order Stiff Ordinary Differential Equations

Bivariate Vieta-Fibonacci and Bivariate Vieta-Lucas Polynomials

ECON 5360 Class Notes GMM

Transmuted Lindley-Geometric Distribution and its Applications

arxiv: v4 [math.nt] 14 Aug 2015

Mu Sequences/Series Solutions National Convention 2014

IFYMB002 Mathematics Business Appendix C Formula Booklet

On Submanifolds of an Almost r-paracontact Riemannian Manifold Endowed with a Quarter Symmetric Metric Connection

Fibonacci Identities as Binomial Sums

Simulation Output Analysis

A Combination of Adaptive and Line Intercept Sampling Applicable in Agricultural and Environmental Studies

Econometric Methods. Review of Estimation

M2S1 - EXERCISES 8: SOLUTIONS

NP!= P. By Liu Ran. Table of Contents. The P versus NP problem is a major unsolved problem in computer

Qualifying Exam Statistical Theory Problem Solutions August 2005

Miin-Jye Wen National Cheng Kung University, City Tainan, Taiwan, R.O.C. Key words: general moment, multivariate survival function, set partition

Application of Calibration Approach for Regression Coefficient Estimation under Two-stage Sampling Design

MEASURES OF DISPERSION

Pr[X (p + t)n] e D KL(p+t p)n.

Transcription:

Global Joural of Pure ad Aled Mathematcs. ISSN 973-768 Volume 2, Number (26),. 95-964 Research Ida Publcatos htt://www.rublcato.com Beta Exoetated Mukherj-Islam Dstrbuto: Mathematcal Study of Dfferet Proertes Sabr Al Sddqu Deartmet of Mathematcs & Sceces, CAAS, Dhofar Uversty, Salalah, Sultaate of Oma. Shradha Dwved Ex-Lecturer Deartmet of Mathematcs, FP, Dhofar Uversty, Salalah, Sultaate of Oma. Peeyush Dwved Busess Studes, Salalah College of Techology, Salalah, Sultaate of Oma. Masood Alam Deartmet of Mathematcs, FP, Sulta Qaboos Uversty, Muscat, Sultaate of Oma. Abstract Mukherj-Islam dscussed a dstrbuto for the study of relablty through falure data, that dstrbuto s useful relablty aalyss. The same dstrbuto s used here to develo, a ew beta exoetated dstrbuto wth the ame Beta-Exoetated Mukherjee-Islam dstrbuto. It s a geeralzed form of the Mukherj-Islam dstrbuto. I ths artcle we have studed dfferet mathematcal roertes detal. We have obtaed quatle, momets, momets geeratg fucto, characterstc fucto of the dstrbuto ad the df of r th order statstcs. We have also obtaed maxmum lkelhood estmates of all arameters of the ew dstrbuto. A relato amog three shae arameters of the dstrbuto s also obtaed, whch gves a vew how oe shae arameter vares accordg to other shae arameter. I the last we have obtaed R ey ad q-etroes to measure the dsorder of the system. Keywords: Beta Exoetated Mukherj-Islam Dstrbuto (BEMI), Beta fucto, Momet geeratg fucto, Characterstc fucto, r th order statstcs, maxmum lkelhood estmato, R ey ad q-etroes.

952 Sabr Al Sddqu et al. Itroducto Mathematcal study of theoretcal robablty dstrbutos has bee a choce of may researchers, a umber of research workers have worked o develog ew dstrbutos to be used dfferet felds of alcatos real lfe data, few of them are, Webull (95), Folks ad Chkara(978), Mukherj ad Islam (983), Sddqu et al. (992, 994, 995), Cha (27), Madel ad Yaakov(2). Eugee et. al (22) oeered the class of beta geerated robablty dstrbutos. Eugee et al (22) used beta dstrbuto wth shae arameter α ad β to develo the beta geerated dstrbuto. The cumulatve dstrbuto fucto (cdf) of a beta geerated radom varable X s defed as, F(x) = G(x) b(t)dt Where b(t) s the robablty desty fucto (df) of the beta radom varable ad F(x) s the cdf of ay radom varable X. Eq. () ca be exressed the form, F(x) = G(x) B(a,b) wa ( w) b dw (2) Where a> ad b > are two addtoal arameters to troduce skewess ad to vary tal weght. Here, B(a, b) = w a ( w) b dw, s a beta fucto. Probablty desty fucto (df) of ths fucto wll be; g(x) f(x) = B(a,b) F(x)a { F(x)} b (3) If x s dscrete, the robablty mass fucto s gve by g(x) = G(x) G(x ) (4) The mortat roerty of ths class s that t s a geeralzato of the dstrbuto of order statstcs for the radom varable X wth cdf F(x) as rovded by Eugee et al. (22) ad Joes (24). Eugee el al. (22) defed ad studed ormal dstrbuto, o the bass of that Beta Frechet (Nadaraja ad Guta, 24), Beta Webull (Famoye et al. 25), Beta-Pareto (Aksete et al. 28), Beta-Brbaum-Sauders (Cordero ad Leomote, 2), ad Beta-Cauchy (Al-Showarbehetal, 22) aeared ths class. We have tred to elarge the class of exoetated dstrbutos ad develoed here a ew dstrbuto wth ts mortat roertes.. Short Detals of Mukherj-Islam Dstrbuto The cdf of Mukherj-Islam Dstrbuto (983) of a radom varable X s gve by, G(x) = x, where, >, x > Or = otherwse (5) Here ad are scale & shae arameters, the df of above cdf (5) wll be, ()

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 953 g(x) = x, Where,, >, x > (6) The relablty fucto R(x) ad hazard rate fucto h(x) of the dstrbuto are; R(x) = ( x ), Where,, >, x >, (7) h(x) = x ( x ), 2. Beta Exoetated Mukherj-Islam Dstrbuto (BEMI) Now we defe the cdf of the ew dstrbuto Beta Exoetated Mukherjee-Islam dstrbuto, F(x) = x B(a,b) wa ( w) b dw Where a, b,, >, ad x >, The df wll be from equ (6) x f(x) = )a x B(a,b) (x ( )b, x > () azard rate fucto wll be, h(x) = f(x) = x x ( a ) ( x b ) F(x) B(a,b)I x [ ] (a,b) where, F(x) = I G(x) (a, b) ad I y (a, b) = B(a,b) wa ( w) b dw The robablty desty fucto equato () does ot volve ay comlcated fucto. If X s a radom varable wth df (), we ca wrte X~BEMI(a, b,, ) We foud BEMI dstrbuto wth same secal case as follows If a = b =, we get back the ma Mukherjee-Islam dstrbuto. Now, comg sectos, we demostrate that BEMI desty fucto ca be exressed as a lear combato of the Exoetated Mukherjee-Islam dstrbuto. Result of above desty fucto s mortat gvg dfferet mathematcal roertes of BEMI dstrbuto. I secto 3 we dscussed some mortat statstcal roertes of BEMI as quatle, momets ad momet geeratg fucto ad the dea of order statstcs. Secto 4 deals wth the maxmum lkelhood estmates of the four arameters. I secto 5 the relatosh amog the arameters s show ad secto 6 deals wth the etroes. (8) (9) ()

954 Sabr Al Sddqu et al 2. Exaso for the Desty Fucto: Here, we wll show the two results of the cdf of BEMI dstrbuto, deedg o two cases: Case : f b s a real o-teger. It follows By Bomal exaso we kow ( z) b = ( ) j ( b j= x F(x) = B(a, b) wa ( w) b dw j ) Where z < Alg equato (2) cdf we get F(x) = (a+b) a j= z j = ( )j b j! (b ) zj (2) x F(x) = B(a, b) wa ( )j ᴦb j! (b ) wj dw j= x F(x) = b B(a, b) ( )j w j (b ) a+j dw ( ) j j= ( ) j a+j j! (b )(a+j) x Where B(a, b) = (ᴦa)(ᴦb), a beta fucto. ᴦ( a+b) Equato (3) shows that cdf of BEMI dstrbuto ca be exressed as a fte weghted sum of MI dstrbuto. Result ca be exressed as F(x) = ᴦ(a + b) ᴦa ( ) j G(x,,, (a + j)) j= Case 2:Ifb s a teger, Alyg bomal exaso cdf of BEMI we get F(X) = b B(a,b) (b j ) ( )j (a+j) (3) j= (4) Now, equato (4) holds the same roerty lke (3), but here sum s fte. If, we follow the same exaso method for df of BEMI we get two results, deeds o two cases, same as above. Whe b s a real o-teger:

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 955 f(x) = x B(a, b) G(x,, (a )) ( )j j! ᴦ(b j) G(x,, j), Where, x >. (5) 2. Whe b s a teger: f(x) = x B(a,b) b j= G(x,, (a )) j= (b ) ( )j G(x,, j), x > (6) j 3. Statstcal Proertes Statstcal roertes clude study of quatle fucto, momets ad mgf of BEMI dstrbuto. 3. Quatle Fucto: Quatle fucto x q s defed as F(x q ) = I x (a, b) [ ] (x q ) BEMI = F (u) 3.. Momets: Momets are used to study the tedecy, dserso, skewess ad kurtoss, whch are the most mortat features ad characterstc of a dstrbuto. If X be a radom varable wth desty fucto x a f(x) = B(a, b) (x ) ( x b ), x > The r th momet of the BEMI dstrbuto s gve by, = By usg bomal exaso, we have, j= = μ r (x) = E(x r ) = x r f(x)dx B(a, b) xr+ ( x a ) ( x b ) B(a, b) xr+ ( x ) a = B(a, b) (b ) ( ) j x r+ ( x a+j j ) dx ( b ) j j= dx ( ) j ( x ) j dx = B(a, b) (b ) ( ) j x r+ ( x a+j x r+(a+j) j ) (a+j ) dx j=

956 Sabr Al Sddqu et al μ r (x) = E(x r ) = ( b ) ( ) j r j r + (a + j), B(a, b) j= Or we ca wrte, μ r (x) = E(x r ) = t j r (7) Where, t j = ( b ) ( ) j r j + (a + j), B(a, b) j= Eq. (7) gves the momets of BEMI dstrbuto. By uttg r= ad r=2 Eq. (7), we ca get the mea ad varace by the followg way, μ = t j ad μ 2= t j 2 where, So mea = t j ad t j = ( b ) ( ) j r j + (a + j), B(a, b) j= Varace = μ 2 (μ ) 2 = t j 2 t j 2 2 = t j 2 ( t j ) Based o the frst four momets of the BEMI dstrbuto, the measures of skewessa(φ) ad Kurtoss K(Φ) of BEMI dstrbuto ca be obtaed as: A(Φ) = μ 3 () 3μ ()μ 2 ()+2μ 3 () [μ 2 () μ 2 3 ()] 2 Ad K(Φ) = μ 4 () 4μ ()μ 3 ()+6μ 2 ()μ 2 () 3μ 4 () [μ 2 () μ 2 ()] 2 3... Momets Geeratg Fucto(mgf) Momet geeratg fucto s a mortat statstcal roerty of a robablty dstrbuto as t rovdes the bass of a alteratve route to aalytcal results comared wth comared to workg drectly wth df or cdf. If some radom varable X has BEMI dstrbuto, the the momet geeratg fucto wll be accordg to the defto of Mgf, as follows: M x (t) = E(e tx ) = e tx f BEMI (x)dx, where,, >, x >, So we wll have Mgf of BEMI dstrbuto M x (t) = E(e tx ) = e tx x a B(a, b) (x ) ( x b ) = (a+j) B(a, b) (b ) ( ) j e tx x (a+j) dx j j=

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 957 = (a+j) B(a,b) j= (b j ) ( ) j [(a + j), t] (a+j) ( t) (a+j) M x (t)= [(a + B(a,b) (b j ) j= ( ) j j), tx)] ( t) (a+j) (8) Whch s the desred Momet geeratg fucto. 3... Characterstc fucto (cf) Φ x (t) = E(e tx ) = e tx f BEMI (x)dx Where,, >, x >, Above defto of characterstc fucto gves us the result as ; Φ x (t)= B(a,b) (b j ) j= ( ) j [(a + j), tx)]( t) (a+j) (9) 3... Dstrbuto of order statstcs Let X, X 2. X be a smle radom samle from BEMI dstrbuto wth df ad cdf gve by Eqs. (9) ad () cosecutvely, x F(x) = B(a, b) wa ( w) b dw x a f(x) = B(a, b) (x ) ( x b ), x > Let X, X 2. X deote the order statstcs obtaed from ths samle. The robablty desty fucto of X r: s gve by f r: (x, ) = [F(x, B(r, r+) )]r [ F(x, )] r f(x, ) (2) Where F(x, ) ad f(x, ) are the cdf ad df of BEMI order statstcs. Here, < F(x, ) < for x >, bomal seres exaso of [ F(x, )] r wll gve us, [ F(x, )] r =( r j ) [F(x, )]j j= ( ) j Puttg ths Eq. (2) we wll get, f r: (x, ) = B(r, r + ) [F(x, )]r+j f(x, ) Substtutg cdf ad df of BEMI to above equato we ca exress k th ordary momet of the r th order statstcs X r: say E(X k r: )as a lear combato of the k th momets of the BEMI dstrbuto wth dfferet shae arameter. Therefore, the measures of skewess ad kurtoss of the dstrbuto of X r: ca be calculated.

958 Sabr Al Sddqu et al 4. Estmato 4. Maxmum Lkelhood Estmato (mle): Let X, X 2. X be a radom samle of sze from BEMI (,, a, b) the lkelhood fucto for the vectors of arameters, ɸ=(,, a, b) ca be wrtte as, L(x, Φ) = f(x, Φ) x a = B(a, b) (x ) ( x ) b = ( x ) [B(a, b)] ( (a ) ) x (a ) ( x b ) Takg log-lkelhood of above for the vector of arameters ɸ=(,, a, b), we get, log L = log log + log (a + b) log a log b (a ) log + log x a + (b ) log ( x ) The log-lkelhood ca be maxmzed ether drectly or by solvg the o-lear lkelhood equatos obtaed by dfferetatg above equato The comoets of the score vectors s gve by, log L = α + (b ) x [ x ] The mle of wll of be, = [ a [ b x ] ] log L x x log[ ] [ x ] = a log + a log x (b ) (23) The mle of, wll be, = d log L da a log a log x +(b ) x log[ x ] [ x ] (2) (22) (24) = Ψ(a + b) Ψ(a) log + log x (25) Ψ(a) = {Ψ(a + b) log } + log x (26) d log L db = Ψ(a + b) Ψ(b) + log ( x (27) Ψ(b) = Ψ(a + b) + log ( x (28) Where Ψ(. )s a dgamma fucto. ) )

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 959 5. Relato Betwee all three Shae Parameters(, a, b) We used three shae arameters BEMI dstrbuto. Relato betwee all three shae arameters, a, b ca be establshed as follows, The df of BEMI dstrbuto s, x a f(x) = B(a, b) (x ) ( x b ), x > Takg logarthm of both sde of above equato of df of BEMI dstrbuto, log f(x) = log + ( ) log x log log B(a, b) +(a )[log x log ] (b )[log x log ] (29) Takg logarthm of both sdes ad equatg t to zero we get, d log L ( ) (a ) (b ) = + = dx x x x = (3) (+a b) Above relato shows how three shae arameters vares accordg to each other. 6. R ey ad q-etroes 6. R ey Etroy: The etroy of a radom varable X s a measure of the ucerta varatos the system. The R ey etroy s defed by, I R (δ) = log I(δ) (3) δ Where I(δ) = f δ (x) dx forδ > adδ R f δ (x) = δ (x ) δ δ(a ) ( ) δ B δ (a, b) (x ) ( x δ(b ) ), x > = gδ (x) * B δ (a,b) G(x)δ(a ) ( G(x)) δ(b ) Alyg the ower seres used before, we obta δ(b ) f δ (x) = gδ (x) B δ (a, b) G(x)δ(a ) ( ) k δ(b ) ( ) (G(x)) k δ(b ) = gδ (x) B δ (a, b) δ(b ) ( )k ( ) (G(x)) k = gδ (x) B δ (a, b) δ(b ) ( )k + k ( ) k k 2 = k = k k k +δ(a ) k ( k k 2 2 + δ(b ) ) (G(x)) k 2

96 Sabr Al Sddqu et al Now, Where, Now, I(δ) = f δ (x) R f δ (x) = v k,k 2 g δ (x)(g(x)) k 2 v k,k 2 = B δ (a, b) ( ) k δ(b ) + k ( ) k k 2 = k = dx = v k,k gδ (x)(g(x)) k2 2 ( k 2 + δ(b ) k 2 ) Sce for BEMI dstrbuto lmt for x vares lke x >. Now uttg values of cdf ad df of BEMI dstrbuto Eq. (32), we get f δ (x) R dx = v k,k 2 f δ (x) dx = v k,k 2 R δ δ (x ) δ ( ) δ B δ (a, b) ( x ) k2 (k2+δ) B δ (a, b) x(k 2+δ) δ dx I(δ) = f δ (x) dx = v k,k 2 R B δ (a, b) (k 2 + δ) δ + Hece R ey etroy reduces to I R (δ) = δ log v k,k 2 δ B δ (a,b) δ Where, v k,k 2 = B δ (a,b) k k ( )k + k ( δ(b ) 2 = = 6.. q-etroy: q-etroy say H q (f) s defed by, δ δ (32) (k 2 +δ) δ+ (33) k ) ( k 2+δ(b ) k 2 ) H q (f) = q log I q(f) Where I q (f) = f q (x) dx for q > ad q R Same lke above Eq. (33), we ca easly obta, H q (f) = q log q q v k,k 2 B q (a, b) (k 2 + δ) q + Where, v k,k 2 = B q (a,b) k k ( )k + k ( q(b ) 2 = = k ) ( k 2+q(b ) k 2 )

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 96 Cocluso I ths aer, we roosed a ew four-arameter Beta Exoetated Mukherjee-Islam (BEMI) dstrbuto. We studed some of ts structural roertes cludg a exaso for the desty fucto ad exlct exressos for the quatle fucto, momets geeratg fucto ad characterstc fucto. We dect a relato betwee three shae arameters to gve a cture of ther varato accordg to each other. The maxmum lkelhood method s emloyed for estmatg the model arameters. Two etroes R ey ad q-etroy are also obtaed. We also obta observed formato matrx. Aedx The elemets of the 4 4 observed formato matrx J(Θ) = {BM m, } = {form, =,, a, b}are gve by, BM = 2 (b ) log ( x BM = a (b ) x log ( x ( x ) BM a = ) ) (b ) (x ) 2 log ( x 2+ ( x 2 ) log x log, BM b = x x [ x 2 ] x + (b ) ( ) 2 ( x ) ( x ) ) log ( x ) ( x ) BM P = a (b ) x log ( x ( x ) x + (b ) ( ) 2 ( x ) ( x ) (b ) (x ) 2 log ( x 2+ ( x 2 ) ) )

962 Sabr Al Sddqu et al BM = a 2 (b )( + ) x +2 ( x ) 2 (b ) (x ( + ) 2 x BM a = BM b =, BM a = log + log x + ( x ) BM a = ; BM aa = Ψ (a + b) Ψ (a) Refereces BM ab = Ψ (a + b) BM b = x log ( x ) BM b = ( x ) BM ba = Ψ (a + b) BM bb = Ψ (a + b) Ψ (b) x + ( x ) [] Aksete, A., Famoye, F. ad Lee, C. :The Beta-Pareto dstrbuto, Taylor& Fracs, (28). [2] Alexeder, C. Cordero, G. M ad Ortega, E. M. M: Geeralsed Beta- Geeralsed dstrbutos, Statstcs & Data Aalyss, Elsever, (22). [3] Cha, J. H, ad M, J. :Study of a Stochastc falure model a radom evromet, Joural of Aled robablty, Vol. 44, No.,. 5-63, (27). [4] Alshawarbeh, E, Lee, C. ad Famoye, F. : The Beta-Cuchy dstrbutos-joural of Probablty ad Statstcal Scece, (22). [5] Famoye, F., Lee, C. ad Olumolade, O. : Joural of Statstcal Theory ad Alcato, (25). [6] Folks, J. F. ad Chkara, R. S. :The verse Gaussa dstrbuto ad tsstatstcal alcato-a Revew, Joural of Royal Stat. Soc., Vol. B4,. 263-289, (978). [7] Gauss M. Cordero, Caral Alexeder Ortega, Saraba: Geeralsed Beta- Geeralsed Dstrbutos, HENELEY Uversty of readg, (2). [8] Joes, M. C. : Famles of dstrbutos arsg from dstrbutos of order statstcs; Srger, Volume 3, Issue, -43, (24). [9] Madel, M. ad Rtov Yaakov. :The accelerated falure tme model uder based samlg, Bometrcs, Vol. 66, No. 4.. 36-38, (2). [] Mukheerj, S. P. ad Islam, A. :A fte rage dstrbuto of falures tmes, Naval Research Logstcs Quarterly, Vol. 3,. 487 49, (983). ) 2

Beta Exoetated Mukherj-Islam Dstrbuto: Mamthematcal Studyo 963 [] Eugee, Ncolas, Carl Lee ad Felx Famoye: Beta Normal Dstrbuto ad ts alcatos, ; Taylor-Fracs Ole, Commucato Statstcs-Theory & Methods, Volume 3, Issue 4, 497-52, (22). [2] Nadarajeh, ad Guta, : The Beta-Frechet dstrbuto: East joural of theoretcal Statstcs, (24). [3] Sddqu, S. A. ad Subharwal, Mash: O Mukherjee Islam falure model, Mcroelectro& Relablty, Vol. 32, No-7,. 923-924, (992). [4] Sddqu, S. A., Deok, N., Guta, S. ad Sabharwal, M. : A ew creasg rate falure model for lfe tme data, Mcroelectro Relablty., Vol. 35., No.,. 9-, (995). [5] Sddqu, S. A., Sabharwal, M, Guta, S. ad Balkrsha: Fte rage survval model, Mcroelectro Relablty., Vol. 34., No. 8,. 377-38, (994). [6] Webull, W. : A statstcal dstrbuto of wde alcablty, Joural of Aled Mechacs, Vol. 8,. 293-297, (95).

964 Sabr Al Sddqu et al Authors Profles: Dr. Sabr Al Sddqu Got hs Ph. D. degree from Agra Uversty, Ida (99), resetly workg as Assstat Professor, Dhofar Uversty, Salalah, Sultaate of Oma. Major cotrbuto develog ew dstrbutos, some work o Bayesa aalyss. Mrs. Shradha Dwved Ex lecturer of Mathematcs, FP, Dhofar Uversty, Salalah, Sultaate of Oma. Dr. Peeyush Dwved Got hs PhD Degree from CMJ Uversty, Ida 23, resetly workg as a lecturer Salalah College of Techology, Salalah, Sultaate of Oma. Masood Alam Lecturer of Mathematcs, FP, Sulta Qaboos Uversty, Muscat, Oma.