Exercise 51. P201: 4.59.Let 54321Y Y Y Y Y <<<<be the order statistics of a random sample of size n from a distribution with p.d.f. ∞<<=-x e x f x 0,)(,zero elsewhere..Show that 21Y Z =and 242Y Y Z -=are independent. Solution: Since ∞<<=-x e x f x 0,)(, so ∞<<-=-x e x F x 0,1)()()()]()][()()][([!1!1!1!5),(424242424,2y f y f y F y F y F y F y y g -=424422))()(1(120y y y y y y e e e e e e --------= Since 12Z Y =，214Z Z Y +=，11101==J . 22111211212112421)1)(1(120))(1(120),(z z z z z z z z z z z z z e e e e ee eee e z z g ---------------=--=)1(20)()]'([)](1)][([!3!1!5)()(11424322211z z e e y f y F y F y F y g z g ---=-== ⎰⎰∞----∞--==01240121222211)1)(1(120),()(dz e e e e dz z z g z g z z z z222211222205042)1(6)1(201120)|5|4()1(120z z z z z z z z e e e e e e ee----∞-∞----=-⋅=----= Thus, =),(21z z g )()(2211z g z g .So 1Z and 2Z are independent.P274: 6.15. Let X be the mean of a random sample of size n from adistribution that is )9,(μN . Find n such that 90.0)11Pr(=+<<-X X μ, approximately.Solution: Since ~X )9,(n N μ,)1,0(~3)(N X n μ-.90.01)3(2)3|3)(Pr(|)1|Pr(|)11Pr(=-Φ=<-=<-=+<<-nn X n X X X μμμ Thus )3(n Φ=0.95, 645.13=nand 35.24≈n . Because n must be an integer, so n=24 or 25.P275: 6.18. Let 121,,,+n n X X X X be a random sample of size 9 from adistribution that is ),(2σμN .(a) If σis known, find the length of a 95 percent confidence interval for if this interval is based on the random variable.(b) If σis unknown, find the expected value of the length of a 95 percent confidence interval for if this interval is based on the random variable S X /)(8μ-. (c) Compare these two answers. Solution:(a) Since ~991∑==i iXX )9,(2σμN ,thus)1,0(~)(9N X σμ-.Pr(|σμ)(9-X |<1.96)=0.9595.0)96.1)(396.1Pr(=<-<-⇔σμX95.0)396.1396.1Pr(=+<<-⇔σμσX X . The 95 percent confidence interval for μis )396.1,396.1(σσ+-X X and the length of it isσσ31.17598≈ (b) Since σ is unknown, then )8(~)(81/t SX n S X T μμ-=--=975.0)P r (95.01)Pr(295.0))(8Pr(=≤⇒=-≤⇒=<-<-b T b T b S X b μFrom TABLE IV of Appendix B, we know b=2.306, thus95.0)8306.28306.2Pr(=+<<-σμσX X . The 95 percent confidence interval for μ is )8306.2,8306.2(σσ+-X X with the length S L 8612.4=⎰∞-Γ===02342121222)4(183612.4))/9((38612.4)(8612.4)(dx e x x S E S E L E xσσσ)29(22)4(183612.422)4(183612.42)4(183612.429402729402274ΓΓ=Γ=Γ=⎰⎰∞-∞-σσσdy e y dx e x y xThen σ49.1)(≈L E(c) From (a) we know the answer is σ31.1, from (b) we know the answer isσ49.1. The two methods yield results that are in substantial agreement, which shows the length of the confidence interval for μ is almost the same with the parameter σ known or unknown.P279: 6.30.Let two independent random samples, each of size 10, from two normal distributions ),(21σμN and ),(22σμN yield 8.4=x ,64.821=s ,6.5=y ,88.722=s . Find a 95 percent confidence interval for.Solution: Let 1021,,X X X and 1021,,Y Y Y denote, respectively, independent random samples from the two distributions ),(21σμN and),(22σμN , then ~10101∑==i iXX )10,(21σμN and ~10101∑==i iYY )10,(22σμN .)5,(~221σμμ--N Y X .)18(~/)1010(222221χσS S +)18(~9)()()()101101(18)(10)()(222121222121t S S Y X S S Y X T +---=++---=μμμμ.101.295.0)Pr(=⇒=<<-b b T b and95.0))(3)()(3)Pr((2221212221=++-<-<+--S S b Y X S S b Y X μμ So the random intervalis )3101.2)(,3101.2)((22212221S S Y X S S Y X ++-+--. Let 8.4=x ,64.821=s ,6.5=y ,88.722=s , we get a 95 percent confidenceinterval (-3.6,2.0).2. Use Splus or R software to compute the mean, standard deviation, skewness and kurtosis of the following dataset.-0.4292, 0.0064, 0.1181, -0.6282, 2.2010, -1.7623, 0.0921,1.8742, 1.4538, 0.3575, -1.5848, 0.4993, 0.7762, -0.2638, -1.1003, -2.2480, 0.5419, -0.4018, -0.3562, -0.5872. Solution: > x<-c(-0.4292,0.0064,0.1181,-0.6282,2.2010,-1.7623,0.0921,1.8742,1.4538,0.3575,-1.5848,0.4993,0.7762,-0.2638,-1.1003,-2.2480,0.5419,-0.4018,-0.3562,-0.5872) > a=mean(x) > a[1] -0.072065 > b=sd(x) > b[1] 1.143067> c=mean(((x-a)/b)^3) > c[1] 0.1234913> d=mean(((x-a)/b)^4)-3 > d[1] -0.5280221So the mean of the dataset is -0.072065, the standard deviation is 1.143067, the skewness is 0.1234913 and the kurtosis is -0.5280221.3*. P203: 4.73. Let n Y Y Y <<< 21be the order statistics of a random sample of size n from the exponential distribution with p.d.f.∞<<=-x e x f x 0,)(,zero elsewhere..(a) Show that 11nY Z =,))(1(122Y Y n Z --=,))(2(233Y Y n Z --=,…1--=n n n Y Y Z , are independent and that each i Z has the exponential distribution.Solution: Since n y y y n e e e n y y y g ---= 21!),,(21, andn n n Z Z n Zn Z Y n Z n Z Y n Z Y ++-+=-+==-21,,1,12121211 and the Jacobian is!1121111021111011101n n n n n n n n J =---=nn z z z z n z n z n z n z nz n n e e e eee y y y g J z z z g ---++-+--+--=== 2121211)1()1(2121),,(||),,(k z n k k n k k e dz dz dz dz z z z g z g -+-∞∞∞∞==⎰⎰⎰⎰ 11012100),,()(So )()()(),,(221121n n n z g z g z g z z z g = and ∞<<=-k z k k z e z g k 0,)( Which shows that n z z z ,,21are independent and that each i z has the exponential distribution.(b) Demonstrate that all linear functions of n Y Y Y ,,,21 , such as ∑ni i Y a 1, canbe expressed as linear functions of independent random variables. Solution:)())(1())(1(11122112211---++---++--+=+++n n n k k k n n Y Y b Y Y k n b Y Y n b nY b Y a Y a Y a n n Z b Z b Z b +++= 2211. Then ,)2()1(,)1(232121a n b n n b a n b n b =---=-- n n n n n k k k a b a b b a k n b k n b ==-=--+---+)1(,)1()2(,,)()1(,111 . Then n n nk k n n a b k n a a b n a a a b n a a a b =+-++=-++=++=,1,,1,322211So1,11+-+==∑∑==i n a a b Z b Y a ni i ni i i ni i i .Since n Z Z Z ,,,21 are independent, thenn n Z b Z b Z b ,,,2211 are independent. So∑=ni i i Y a 1can be expressed as linearfunctions of independent random variables.P275: 6.19.Let 121,,,+n n X X X X be a random sample of size n+1, n>1, from a distribution that is ),(2σμN . Let n X X ni /1∑= andn X X S ni /)(212-=∑. Find the constant c so that the statisticσ1+-n X X chas a t-distribution. If n=8, determine k suchthat 80.0)Pr(=+<<-kS X kS X μ. The observed interval is ),(ks x ks x +- often called an 80 percent prediction interval for 9X . Solution: )/,(~2n N X σμ，),(~21σμN X n +，),1,0(~21σnn N X X n +-+ ).1(~11)1(1),1(~).1,0(~112212221--⋅+⋅-=-+--+-+++n t SX X n nn n n nS n n X X n nS N nn X X n n n σσχσσTherefore 11+-=n n c . 80.0)9797Pr()Pr()Pr(999=<-=<-=+<<-k S X X k S X X kS X x kS X 60.1,415.197==k k . So k=1.60.。