Scribd
Upload
7 views65 pages

Statistics PDF

This document discusses continuous univariate probability distributions, focusing on the definition and properties of continuous random variables and their sample spaces. It highlights the differences between continuous and discrete random variables, particularly in terms of probability assignments and distribution functions. Additionally, it covers concepts such as probability density functions, cumulative distribution functions, and the calculation of moments and expectations for continuous random variables.

Uploaded by

demo.user2025
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
7 views65 pages

Statistics PDF

This document discusses continuous univariate probability distributions, focusing on the definition and properties of continuous random variables and their sample spaces. It highlights the differences between continuous and discrete random variables, particularly in terms of probability assignments and distribution functions. Additionally, it covers concepts such as probability density functions, cumulative distribution functions, and the calculation of moments and expectations for continuous random variables.

Uploaded by

demo.user2025
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
Cuaprer. 19762- CONTINUOUS UNIVARIATE PROBABILITY DISTRIBUTIONS 1.0 INTRODUCTION . bi We have seen that, there are three types of sample spaces (i) finite, (ii) conutably infinite and (ii) continuous. In this chapter we discuss the random variables defined on continuous sample space (A sample space which is finite or countably infinite is called as denumerable or countable. If the sample space is not countable then it is called continuous) In other words for a continuous sample space © we can not have one to one correspondence between 2 and set of natural numbers (1, 2, ..... }. Illustrations of uncountable sample space : (1) Suppose, weight of an oil bag having the capacity of 1 kg filled by an automatic filling machine is noted. The sample space will be an interval in the neighbourhood of 1 kg such as Q = (0.980, 1005). (2) Suppose in an experiment, life of an electronic component in hours is recorded. The sample space in this case may be an interval as a part of Rt such as 2 = (0, 5000) Note: A continuous sample space is a subset of real line. 1,1 CONTINUOUS RANDOM VARIABLE {In general, we define a random variable X (w) as a real valued function on domain Q. If the range set of X (w) is continuous the r.v. is continuous. The range set will be a subset of real line! “ Illustrations of continuous r.v. (1) Weight of a person in kg. (2) Consumption of electricity of a town in a specific month. (3) Daily rainfall in cm. at a particular place. (4) Instrumental error (measured in suitable units) in the measurement. (5) Life in hours of an electrical component. Note : The distinction between continuous random variable and discrete random variable is as follows : . __ (1) A continuous r.v. takes all possible values in a range set. The set is in the form of interval. On the other hand discrete r.v. takes only specific or isolated values, (2) Since, a continuous r.v. takes uncouniably infinite values no probability mass can be attached to a particular value of r.v. X. Therefore, P(X = x) = 0 for all x. However in case of a discrete r.v., Probability mass is attached to individual values taken by rv. In case of continuous r.v. probability is attached to an interval which is a subset of R. (1.1) HNjOS 1) rarpt09 Saxe Ton fm end One 38th) = 8 Ci) ta) = Hy 20 imptat 2 0 free i ltan-? col” = tan kway) = 1 kre kei ‘From (a) and (b) we get k = Coe aies Bx.1: Suppose a continuous r¥. X has pdf. to=E ; -1sxs2 =0 5 otherwise A « (eix20} iel-}sxsi : Be eldest), fad P0A)POD,POA), AB), P AUB) PLA'AB), PAIL BY, Pian By ie 2 a 2 gaz Jena = [Pace wt = 00 ii) 1 z fi Post dexs}). [ewan - Loans el wes ain PUA) = 1-PW «1-89 «10 ww) Pap = Posxsia« J toyar= f ¥ ae 3 2 pit =p bb? eee eraeeeeteeeeeceeieeeeeeseeee % can ui. rb. sent = —— one (AUB) = P(A)+ PCB)~PLANB) = P(-V2sX52). oe : P(-v2sXs2) come “Sta PUB) = PB)-PIANB) = PC UR tbat f ) OR CUMULATIVE DISTRIBUTION FUNCTION 1.3 DISTRIBUTION FUNCTION (D1 cor) isuiote rvs The summation ie to he replaced by integration, ‘Distribution function is an important entity in the fleld of statistical inferene: reliability theory and life testing ete Definition : Let X be a continuous rv. with pdf. f (0. The distribution function « cumulative distribution function denoted by F (x) is defined Fix) = P(X Sx) ; -wexce = [tee Note: Ar. Xia defined tobe continuous if (is continuous Mstraion 8: Av. X has pat tw =2e", 350 a Find is distribution function Solution: By definition, Fave [riod = faettae [EST eroee ge? Sine Fe defined fr allx« Re hav owt Fi =o x<0 : = 1-te® | x20 ‘ nn ENE enna ee g 8-O=1 3 and (b) we get k = 9/8, therefore f(x) = 2 (xp k Tox? 20 implies that k 2 0 J tora team! wo) : 5 i ¢ E k [w2 - (— 22)) kre k =n From (a) and (b) we get k = bywe getk = 2 tia fo = 1, ie Ex.1: Suppose a continuous rv. X has pdf, “3 A= txix20}, find P(A), P(B), P(A, PAK P AGB), PCAOB), P (AU BY), P(A‘ B’), Setutions i) Pia) = Poreo = J rwrae = [Saxe Lif = on Gi Pcp) = p(-bex tat f fo dee 1.3 DISTRIBUTION FUNCTION (D.F.) OR CUMULATIVE DISTRIBUTION FUNCTION A distribution function is defined in case of a continuous rv. analogus to that of discrete rv. The summation is to be replaced by integration. theory and life testing ete Weflnition : Let X be a continuous rv. with pat. cumulative distribution function denoted by F (x) is defined Fa = PKs») = froma = ee defined tobe continuous if F AMustration 3: Arv.X has pat {a= 20m =o otherwine Find its distribution function, Solution : By definition, ‘The distribution function or Note: Arv.X is continuous, x>0 ewe aha a <2 y= ftwa = frees [SS = 1a ge Since, F(a) i defined forall x © Rwe have to write, Fin =o 1x0 = 1-Be®, x20 \ ee ¢f as follows NEO the BAL. We can, fing anaes P=) Par sy8.e (Sem. 4 fa 2 rw stributon Rincon of rv. X having pa et Ul tan = Berbeos! wenn (2) Given the 4 Fin we can find Pia €X £6) a flows toelet 5 es PIAcXsb) = Puksm-Pxee, soho tieb | le » Properties of Disrbution Pumctio x fiese (1) Now-negative : Fini money oxen rae from. fy = Pxewe 0 a Tianhe baa F (x) being a probability it is always non-negative, = glen a gk ® Now Mt # Fx)iswon-decreanng The Las Proof: Let, A= (X| xa) and * Be last) clearlyaca cased nes 7 Pie ray t Fone fra + fea Paxse = Pxe - Fas Fo Peo and Fim) = Bm py Note that, Fim) epaxe F= Pasa sy fyeva ft etog ©. since tx | xs XSlisacure event 20-063 Leena] oad Fons dees x tbe os MEN: Goaphot pee = 2 oH Pig LAs Grapher Bx. 3: Ifthe distribution function ofa 1-v.X. is, “as Fo «2 fan d+ 21, find its pdt Solution : Pa fis given by. con_un. Prob Daibtns 7 ae ce ra ih 9a 0 Saitae 185 on guano pal ( hn the eae ai oy = 22 tan oo 22) fay = 22 tant : feces po frtwe fast as ciate | neence ‘ - aioe x].ft0) dx), ‘Ex. 4: Ifthe distribution function ofa rv. xis jded the integral exists (L€ fi i fee xe fe ) as a real valued funetion of 9.x, Then B [g(x] is given by, =R j Osxer oe = M-wR-1 5 Isacz : z to Foo de cc a) find is pA £19) Alo draw the graph of f(x) and F ( : Seaton: 0 reo Frets [can anieren i oe | mtn teed hereto fad ig) we need not find pf. of, using pa of X we fx; ire can nd Ie Oe o wee | atsiration 4 Suppose, the life ofa electronic component in hours is a continuows Thus, oe ee Mint pa Bx 5 exer | Fy = 7002100 ee eee Find whe expected life in hours ofthe electronic component | swans oo Patina =f 28 gy : | : th i: s 20000 J x dx © 20000 [- _ = 2ooonf oy q y | i 1 1 ina nom negat x AstiVe FY. then show th Mies 1.4 EXPECTATION oF Fig 10 | vet on CONTINUOUS RANDOM VaRiABLE onal a eae NO aed We have men how Ho Ji won 1 Hotution | * Integrating by parta we get, learn the concep tua ; br canna “Yt of eer 2 Jn-wee i 2 Th simatic On inne it eed ; eee ites inant MEN ed af Gt i MT fxdurenae J xe tana C Fmen + [etn teen Prints - Boer sa Saxtiidee nan. Jets = anon nn w 8 fetevtese frre : = EoD - Penen eOT) = EH * «sad d somate, We ssa tro mer teal rewarding bivarate t% shich wil be proved ater on moomem 9 : IX and Y are any ws continua ry. then, G) BOLE » kousEcn 1) KEW 60) « ABO» MEO 66 ss “SYBE: (Sem y Santee un cee aie Ex) EM fi chh tn ereer onan Fw a(h)eat on eR) eo Wye BB BP oes ace ecto ly iad tifa ew lle donated by and ise defined a, noms fat tude provided, J ti toaree ‘Therefore, the fst four raw moments wil be, wie Pete onan aye Perna soe Pwtioe Central momenta: IF 00) = m then centeal moment is given by he = BO mee fixe mi tine provided, Lum mF | fon dec Coan, Me = Em = 0, iy = BK -me = Varco, Ms Bm) and jy = BOK mit tral moments and raw moments wane (ai)? elation between cen OND TH Hs 04 G)" (uh (©) Moments about «wil be ys a) < beer | nec Ma ns heat Bd yn Harmonic mean (i) is roy & ven by, Ho Q)> PF coo Muatration 5 UX isev. with par Tesla our raw mente of Sioa tral moment of a we ‘coefficients f,, p,, Hs Ye: “ Solution: (i Raw moments eye mel WOO = Gateoaee 6 feta nae format a] m= 800 ~ Prerwde ef eta ade x] eg.3 5] a =o tas on. iw, Prob, Distbuons - 5-5] - 8-4 HL = Boe - ehh GG) 1a.e 3 4-288. «iy Presi /ih = 0 and n= VB = 0 2 e Bee mid = gy x. 8 nemaeBag Inui (1) ice diate payee = 0. the distribution in symmetric. (2) yp < 0, hence the 1.6 MOMENT GENERATING FUNCTION (§.0..) Te coma a many ways in the study of probably ditibution sch, ) to ind momenta 1) (0 Snd probability distribution of gx); function of rv. X. satis gan tether the random variable ae independent in cae af multivariate Probability distribution, irae Arana wth phen momen generating faction Xie ented by Mgt ea eae Ma = Bat = Fecrarar Provided the integral exat for some h such that -h- 0 To ae Mie ‘Baw moments using [Link]. pine 4 = PP raw moment + etisent ot [at Mee) Method 2 ad ce Generating function for central moments EX = Meomit = Efe] cata moment lows se + coetineto [Mem (ar he Nove :06F tr moments bat canbe Bluneion 6 UX wars withpar ne te™ x30 find moment gener Also id 8, 8 otherwise ‘and interpret, Solution: By definition, = Beef etraae cafe fee 2 By Away Expanding My (a8 power series int we get, (1 eay ; : pops : My = coefficient of ¢ = 1 2 ee ‘Note: Mt) san be expressed as power seres in t a8 follows chi Pte Ee tn the expansion of My _ lefined similarly Tt is My ing function. Hence, find the =n ne dn the expansion of, 4, -MGP fu central moment wil be, © E(e) =m My mtb (Integral exists if(2 -1) >0iet craalff en ie 1 = sent fe = 4 (3) cant momenta: ea ome Oe aE wy -3u + 20" Bee Baa 8a OP = 8 sf bd@p ag) 3 3,32 =$-2,2-2. ons a7 ata7 we > b= wiv as = 122,38 ad = nla se Interpretation : (1) Since, 8; # 0 and y > 6 the distribution is positively skew. (2) Since, = 9>3, the distribution is leptokurte. 17 PROPERTIES OF M.G.F. 4) Me(0) = 1 (2). Elfot of change of origin and scale Renult: If isa rv. with MG F. Mgt) thea, (1) Meat) = My(t) ett (10) Mey) = Mgt) CH) Meee) © et Mg et, where a,c are constants POE Aen = Ble =e |e Bfete et] cat ate “i Merit) = B[e*] o Bfet)] = ayien cu Mesa = B[oeene] . pfateas] = Ee). oat a 6 a and Y are independent rv. then, Mawy(t) = Mg (0) Myo) aa te wer series int a8 follows Bite: My) canbe exprned per sere () = Bie) = Baste My ts # My (t= 14m tomar +H Raw moments using MF. Method 1 = #8 raw moment = coofcient off inthe expansion of Mt, () a My (07) Method: Wee tara _Senerating function fr central moments EQ) « m then MF. for central moment wil be, My it) = B[o!-m] tm p(x) gives central moment jy a6 fellows, Me) Me = content of! in the expansion of My — y(t) : ternative [a Mees (W] oe a ‘Note: [Link] for moments about‘ can be defined similarly. 1 ia Myatt) ety, Plosteation 6 : Xin ary. with pdt ee a =o otherwise Abaca rating function. Hence, Gad the frst fer raw and central mons Also find 8, Bs and interpret, Solution :By definition, BO Be [etwas fimens + afetetng « feevey or i teh i (Caegral exist if 2-0) > 040 ote = avy! erudite serie nn yu i Baer. Oar wh EY... (0a fff ey My = confileat ot t= 1 weno me 20) y.3 3. B= BA 6 oud IE 3 Be and = Intorpretation : (1) Since, 8, # 0 and uy > 6 the (® Since, b, = 9>3, the distribution is eptokurte 1 PROPERTIES OF M.G.F. 4) M0) = 1 42) Effect of change of origin and scale. Result: IX isary. with MG. Mee) Ben, (8) Mesut) = Myit) ett (1D Mex (= Myce) Ci) Ma sex (0) = ett By (, where a,c are constants Proof: (iy Mea = 8 [eho] eft ot] cet mg wy Met = Ble] we feste) wm ce aiy Maser) = B[elecoe] . p feats cx] + AE [oreo]. tae, (ay istribution is positively skew. ol tae caieecee| Por a given pf. there is unique o- aC ysndhe scope of the book. Though, we are not proving the Pe or tea estar ma, ty diembtion using thie property we can conclude that the and central es P- 3 sts TMi) = a= 48 Prot: Mery four eurmula nt Y on eta ca a x = coefficient of $= 3 ot as ! ei and pit eo a a = coatint ot f= fe “ foe. e 4 i 4 j ei eee t weet 3 perties of CGF.: : isu (9) aan bare snstate hen, 1.8 CUMULANT GENERATING FUNCTION (€.6.F) Pee a teae 0 order to obtain M.G.P. we have used the transformation ot8, what do. ees X and Y are independent random variab Use inverse transformation be 8 question of interest. Accordingly, (2) 1X and Y are independ ghtsined. 18 called as cumulant generating funeton, which ia foun Kes vi = Ke(+ yo cae ae MEDIAN AND QUARTILES oo Pefcition SIX isa. with MOF. My (than loge Me () in elle HX isa Dh pa £00) th sendrating function It denoted by Ky thecurve y = f(x) into two equal parts. Therefore, median (4) is.a value of rv. X that Thos, Kei) «og Me). Like MG, o also be expressed POXsw) = POX2w = 2 K, + he tH! + ey C881 t F 7 [ar Kec i ky = | F : os . tmulants and momenta Key b= es ba ety, ky ‘Thus, we get first four central mon, Hs = 0, a = ke, Hy = by, 1 Fwe} Foie 8 of a follows, Po We ksau : istribution function, ies of rv. X or probability distribution can be defined on similar lines Quartile 3 foe awe rue ee iL Fes} i tn ea nome Pesaran eta gnrt Ghat. X aud a, j pacay+3 © Pare =! : ree! rh srt bcd ele tia nal ail sod percentiles. Dacilee: The decile (D1 PasB)= 0, 4 © Fm) auto Posh) = x00 2.98, Mustration 9: WXis ac with pat fa) = 0%, 250, 650 nd the quartiles of X and quartile deviation of X. Soluion: ingrniy Pas « | Habae value of X auch that, X such tha, oo ery Qe» — tage 1-4) 1 = Fe i = et «tg and Spe) dog site (= then neste £608 MES tne ate ip is» aerennbe Nm ‘dx fllowing conditions sax which sais the f fa In order to maimine £() woolve £2) = 0 oes loge (0) = “log + Boge x iorstiting both the sides Wirt x, we get, cats fo) "= fo GJ = Oimptien 1) 0 ie x =2 take c= 8@-).t0 ea = Bi) tea 9 an faker =0-} 10 X~ 2a point of maxima, Hence itis mode. - Note: Sometimes mode can be obtained without a itty A ote ot ke ‘4 pA, then obviously £ (x) is maximum, whenever (x ~ 5) is 1 tmareaprinitum value of ~ 5) is zero it is atx = 5. Hence x 6 is a mode (1A) TRANSFORMATION OF RANDOM VARIABLE sing calculus method. For instance, @ Guarie Deviation = 8-9 ae elie, we vars pat We tt intrested inthe probably ditribution of function of «random occas ven the probability datributon of X we may be interested inthe 1 = ype fditribution of ax FAK +X, IX], oF, log X ote. It may be useful in finding intereletionship pe nunons Mainly, there are three met “nation of, function of rv. as listed below thods of obtains, nm a REE Susstormation method, which i tobe used for strictly mx 7 one-to-one on*e function, enetene g 4) Using distribution function 4S) Using MG.F. of function of rx According to the ype of function the method is used, Method (1) + Suppose, X is a F (x). Suppose, Y= g (x) is strictly 1 given by. tious With PALF and dition monotonic one-toane, onto funtion, Tye vars so = rn |e Since. Y= (x) ts a strictly monotonic onto function, X= EOD ie Wel defined ey, bin = tetany Mlustration 11: 1fXisary. with pat She fd tenet ot Y= Solution: Step) Pinter be. x Here ¥ seo X vtex sy poly 1 X x ox sere vot [55 » [Ee ‘Step (3) Find the range of Y. IF. ‘monotonically increasing function, ‘Step (4) Find the pa of Y m0 «0, [Rl = egy | = OY = OandifX = 1,¥ = = Note that, Y= bie jee y 9(Cex) aap + 0c en ye =a 1 On ¥ew |578Se (Som sates (1) Son un, = X,¢00 = |X}, then Method (2): If ¢() 6 non-mentonic function, such as g X) = X2,g00 = | (isnot well defined, I due tothe reason that 0 is not one tov fanction, I thi rmethod (2) fila. To overcome this difficulty we find HCY) the distribution function of cue, ‘s YeHence weet tbe pat HOD = Ht Fx 611K ie = IX i ‘Solution: (1) Let us find H() the distribution function of Y= X? HO) = POsy) = Patsy) = PCAy ex sv) taking values in (- has pAf F(X) Bnd PAL of G) Y= x, Ho = POPU, ocysat ALO) 2 8 Pas yt ay POD - Ze 1 5) 07 = 105). te) cy, bon = t0).51e tea ge 5 mor Sg EGS) tC), orocy eat Note: IY iesymmet around 0 then, £05) = mano = Ey (18) Lat G be the distribution function af'2 = |X} Gin) = Paes) ~ PuX} en, Gi) = Persxen Ostsa G@ = Pa-PeD, Thep. af ofzis © Gy PAL ofzis 2 Oc a 4Fa dren aS = a a B® © f)—fe9x ey £@) = fe) ston + if0sese Notes If) aymmetrc then (3) = £(-3) Hence, (2) = afte tay Sy =o, -1exe. Ad pat oft ¥ = x8, iy Y = |x] Solution: W) Patoty = hye 13, V2 3 ¥ then BO) 0... BO, 24. Probability density function (pd), 1a) You) Yes Pia) = 04 Pane) 18. 4) = Be wt) Pw Gi) ey) = 2 a 4 221 a. is 25. (EOP. 2. Cov Y= 25 29. LM. a raw moment pan Pian’ 128 UP ANSWERS. NSWERY Yes (0 Yes (@)No (hy Yo wp, -lexel x<0 Dexe1 Oeyer 26, Continuous © 30. Aatilog 1 oe 32. BOO 2 EO 4 sacciate two real valued inspection of spring is ach spring. Thus, with jointly variable quantities. The joint bation to be introduced. Moreover, oor more random variables. In this oovcTion ae ge 2 is a uncountably infinite sample space We can ai or example, in a manufacturing proces® aswell as tensile strength of © | juch variables needs bivarite probability dist Been to find the distribution of sum of ts se iy study bivariate continuous type of probability distributions [1 CONTINUOUS BIVARIATE RANDOM VARIABLE Definition : Let, Q be the sample space corresponding to a random experiment. Kor and ¥ (w) be two real valued continuous functions with domain 2. Then the ed pair (X (a) ¥ (o)) is called a bivariate or two-dimentional continuous random be, being sample point in 2 [Note (1) For simplicity we write X and Y in place of X (w) and Y (a) respectively (2) IFX hae range set Acc Rand Y has range act BR then, &Y Os xB, lustrations : (1) Xt) = Weight of ul student. ¥ (a) = Height ofat® student (2) X(w) = Distance ofa space denoted by w from equator. Y(@) = Average temperature atthe same place w. (3) X(@) = pH value (acidity) of water ata pond at time w. ¥ (©) = Growth of fungal plant at the same time © JOINT PROBABILITY DENSITY FUNCTION lier in 1.2 we have seen that in case Pam: 7 of continuous random variable, probability i Machel any Sndividual value. 1 (X,Y) n'a bivariate random variable then ihaniahed to 4 two dimensional region, wach as (9) | @ 0 IBS Pinyet = Flay, for xfsed and k> 0 5) Macbandc0 = fern bly = bey ie : ee tlh = ty; oeyer Solan: Narnal pdf Xen by no) = fe Fo nga [ter deve ; fy oo X20 t lacy Deyel inl 4 oY i given by Hence X and ate depondent ! a EXPECTATION, MEAN AND MOMENTS Ae Definition : 1%, ¥) isa bivariate rv. with joint pat £0, Y) and g OK, ¥) ia a + [er daetf et dcvet sy50 on X,Y! then expectation of g (X,Y) is dented by B fg (X, Yi] and ies defined ae Note that f(x,y) = e-(x oe FOP cer ey = He). GORY @Y). oe X and Y are independent ry i fia Provided the double integral exista Note: [Link] find pf of = g CX, ¥)and then find Ee Suppove (X,Y) is bivariate rv. with pat Lig S10 otherwise abide whore h @) is pat of ‘Are X and ¥ independ, bate en To get raw moment af order 8) we take (K,Y) A= (yy) locxe1,-y ie : ffx se som. 8 ere in partial Z =Ew a ee for 1, wert Hy = BOD for weet, = BOY) ibe eM meeser: r 2) eee Bac) fo oo = EO) vermene (OY ov 8, ¥1 © Hhy~ os Phe central Moments: If (X) = m, E(0) = ms then the e (y-m)* us = Elem Jox-myF cr -my)* £0, tao = Var ©) bon = Var CX) + tin = Cov OX) coeficient of correlation is, Cov (XY) War ) Var (Y) Via Mlustration 9: Suppose X, Y)is a bivariate rv. with pds 2 Dexeyer otherwise Find Cone oxy Solution : T» find means and nat reans ond variances of X and Y we Therange A ( (11 B04 gehay, | 9 OM. (0, ty Part (2): Here we ass, May tit) = My 4, Note that, int M5704) and we want to provethat X and Y | Mey) « Me .0 Mx, y (0, t) Jn itp sims gyese (com tL G foapdrdy = fot® goods fee Roray fewer =f fe rerpatoraa, £9) = §0).80),¥ ey) ‘Thus X and Y are independent random variables, Mlustration 10 : IFGK,Y) gary. with pat faEM=eEM | yoo, yoo =o + otherwise find M.G.F. of X, M.G.P. of ¥. Verify whether X and ¥ Find M.G-F. of &, ¥), Hence, sites hyue FPS oH ay 81-8 7 a MOF. of X= Mey (&.0) and MG. of ¥ = ay y (0,4) May (sta) = My, y (1,0). Mey (0, ty, X and ¥ are independent rvs, istribution of 's using conditional probability distebericn ot ous Bwvanate Ostnbuons Sess Pt continu rps en AES rate r¥. with oir PAL, 9). Sop 221 aie ee Definition Let X.Y) * 1 saitinal expectation of X given y se etait enditional pat of X 600 F Dee eye Ep and ts dened a : To; otherwise a ional distribution jtional distribution of X given y, (i) candi x gy) x | 23 oS On vvvditional distributions we need 19 ind Se [ sotution : Ty, Caprese the range set A suitably £ Fe a — sete jocreyell, Ars (ap locxehxey 7 : : Wevan obtain expectation a flows | i EEA) = EY f | Conditional distribution of Y given x is Correlation coefficient and conditional expectation : S: a isa linear function of y then the coet + regression coeficient of Xo Y is b othervice , 1. Define continuous bi reewole) 2 Dt tain rite dtnban Defie joint probability ati ' iti ution funetion and distribution fonts | Waite 5. State the i Properties of dit Bopha tore bution function of a continuous bivariate | co cept of margin ot nestion with ‘ginal and conditional probability distrib” s) se ‘continuous Bivariate Ostbutlon® aan se MG. of, 1 whee 9 Ha comtinnos bivariate rv.s em suatics 1 pagal netation show thay Morey ett) = My (abMy (< ee Mca (t) and My ee ‘and that of Y given x. If the 1 expectation of X given ¥ 7 a ‘hen interpret the result. Also, Dein conadipetations are inet fonctions odin pee terclton In ease contol expectations Fe eer | fnstons. {unser independent rcs. then, with usual notation show that (BAY) = BOO EC {i My ict) = Me td My ae aes Ya abivarae wih it tf 92. Suppose Un Us 9) and Sonne ier and, which ate uniquely snverible such that 3 Bookie how to find the marginal pat of X and X= y(U, Vand ¥ = 00, thot of ¥. Make suitable assumption if you need {continuous rv with joint pdt. f(x, y) for 160%, Y) in & pair of independes £5 0.y>0; show that the pf. of Wy UeSin nee J tarnanee ain Vex, pw ft (sa Hint : consider i) U=X-4¥, V=X, (i) U= YK, V= X, (ii) Us XY, Vs X and use transformation method, Express each ofthe following regions of integration so that 4) the limits of X are constants (free from y) Gi) the limits of ¥ are constants. Mxyix>0, y>0, x+ysad () [wylx>0, y>0, xey) (6) (xy 1x20, y>0, x22y) 1 (xy 1x50, y>0, xteyesal (©) (xylx>0, yoo, xey2a) (f(xy l osyy I of X and MGR. oY = mean and median of X- pdf of X and Y is ev), PaYSx istributions[Mumbal University 1997] A inal and conditional distribution Ker. with joint pal ea deeay OSnySL VP pn distribution of X and that of Y ssa deribtion of ven y and that of ver a, iz a iv xy, ob aX, a 21 oyun pa of We aE HEY, genen, -xeyex eQ[Link] ¥ are independent © omm5 Gy “0.4095 a) (0004 Gi). 102 © hae BEE Ww Gey = 28g © trey 0 223 _ tes ey 9) iy = Sw gon = 228 6) Gan = 2m; oemer fii) 0 Norma jae [fae Mame 8 y Normal oistibution aves (P-") a3 oe x 02, (oe, (2 8 even function) then, 2 eee — : wae eer Jom -2fe oven = « (x), Fe. 00 is odd function, then, - elt Y fama so 2 Boo i (-b-*) Substituting, y = "5 4. The probability density curve of N (u, 0%) is bell-shaped, symmetric about y as EF A 7 ‘esac sha inthe fling gure. Hence, prababieydeneio rere cea ‘metric abot ree Bao Note that f Hy We = Ae eae fuxstion Fig 3.1 Pig. Ve peer ‘5. The maximum height of, probability density curve = —1L Var 00 = BX-EO0 =p o Van el 3.2 MEAN AND VARIANCE oF Nua) In the course of derivation of mean and variance and moments, we rea" aoc c Sow : Jt ae Puy one bygones oV2x - fttsy 2 ove | the integrand is odd function.) Gigi ee Trial oda ordered central moments of N , o2) a ntl (1, 6 are zero, i J ewe Sew Se » fem == ova '* Normal Disrbston function, we get, Hence, normal distribution is symmetric around j. From pdf. also we got that, fu-a) = fuse, va ‘Thus pat ie a symmetric function around Karts Hence, normal distribution is mesokurtc. ‘Recurrence Relation Betwoen Even Ordered Central Moments: Statement: Ifygr is a central moment of N (i, 0°) the Mar = Qr-V0% Yen Proof: Note that, ar = E(K-—iir . wot feat vel 2 2 “ coefficient of tin Ky(t) = = mean i i ka & coefficient of fin Kalt) =o = variance content of in Rl = * ¥ ft) =0 ur © Or Da 344 [Link] CGF. OFN Wo") MGF = Myit)=Bet 0 for 23 Here west, Kr ° ote MGR, of standard normal variate ise” and [Link]. is 5 4, isTIBUTION OF LINEAR TRANSFORMATION OF NORMAL YARIATE ION oF to use the probability distribution of linear transformation of bation can be obtained by directly deriving the pd.f. or by 1 -f ee Putye Ex ° Frequent oonal variate 108 distr feowing GF Result: Let XN (, Proof: LY MG. of ¥ then aX +b +N (au +b, a? o®) where a and b are constants. ax+b Mit) = Max op (®) PD (at) ce plate Fottant ag obit hat ote? = ME. of N (ay +b, 2 Thus by uniqueness theorem, we ean say that, aX +b Nau +b, 220 “env on (5 ¢ b x Therefore MEH Sf +N (0, 1) Thus, we can convert any normal variate to standard patsy 36 AREA UNDER NORMAL PROBABILITY CURVE » he density Try Earve F(a), bounded by X axis and the ordnaies X “+N (yo then, WX w Px>a) ew nin a ; sated directly by using weval ‘ani he tga cane be we ec ant rah Me TS omy teal able The cca Sc, aig efrmain = ~ rote igh sie fr 0, DL. POLE A) «> 0. In “en ewe the long rls rt, nists 8) —— ‘The able ies the probability of any P pettus dna FH 2, The pd curve N (0,1) is symmetric about 0, therefore if we fd the about ¥ ani ie at X= 0) the two part coincide. Thus, we get area of right ta pametric ef tail same. In eter words fr Z-+ N(0, 1), PZ >a)= P(Ze a) can be Coa) vr ee ‘pce can be also computed by the fllowing form ipoos if 290 or x>H sed ay ° tegral table tml oa : ila with « fair degree of | Saeco Be oot9s27 196864, ¢:= 0.115184, c= 0.000844, c ‘th above approximation can be used to write @ computer program to evaluate (2) or Figs ze 3. Area below the ordinate = 1 - Ares above the ordinate q Paca 1-Pase) Disrtion ntion of 4.09 Quite often we re ux) ocan be computed using calevlator. putation of Probabilities XN, then, i svire to compute distribution function of « normal variate © then the distribution funtion Fix is given by, E fa Pacten = p(t2teKit<2ct) : 034) -o(3#) ilustration 1: 1€X + N (3, find () P CX > 6), (i) P CX < 1) Gi) P OK > 0, X=), ) PB- 4), (rtd PCIX~3| < 3.92) We denote dateibtion ation of ¥( Hence, : Liby 646) $0)<05, if 260 w Xep

You might also like