STAT542 HW10 SOLUTION 1 4.63 Since X = eZ and g(z) = ez is convex, by Jensens InequalityEX = Eg(Z) ≥ g(EZ) = e0 = 1. In fact, there is equality in Jensen’s Inequality if and only if there is an interval I wih P (Z ∈ I) = 1 and g(z) is linear on I. But ez is linear on an interval only if the interval is a singlepoint. So EX > 1, unless P (Z = EZ = 0) = 1. 5.3 Note that P Yi ∼ Bernoulli with pi = P (Xi ≥ µ) = 1 − F (µ) for each i. Since the Yi s are iid Bernoulli, ni=1 Yi ∼ binomial(n, 1 − F (µ)). 5.11 Let g(s) = s2 . Since g(.) is a convex function, we know from Jensens inequality thatEg(S) ≥ g(ES), which implies σ 2 = ES 2 ≥ (ES)2 . Taking square roots, σ ≥ ES. It is clear that the inequality will be strict unless there is an interval I such that g is linear on I and P (X ∈ I) = 1. Since s2 is linear only on single points,we have ET 2 > (ET )2 for any random variable T , unless P (T = ET ) = 1. 5.12 E Y¯1 = E X̄ Z ∞ nx2 1 e− 2 dx since X̄ ∼ N (0, 1/n) |x| p = 2π/n −∞ Z ∞ nx2 1 = 2 xp e− 2 dx 2π/n 0 r 2 = nπ Z ∞ 1 − x2i E |Xi | = |xi | √ e 2 dxi 2π −∞ Z ∞ x2 1 i xi e− 2 dxi = √ ·2 2π 0 r 2 = π r n X 1 2 E Y¯2 = E |Xi | = n 1 π Thus E Y¯1 < E Y¯2 . 2 5.13 ! r σ2 S 2 (n − 1) c E n−1 σ2 r Z ∞ n−1 σ2 1 √ c q n−1 (n−1)/2 q 2 −1 e−q/2 dq n−1 0 Γ( 2 )2 p S 2 (n − 1)/σ 2 is the square root of a χ2n−1 random variable r Z ∞ Γ( n2 )2n/2 n σ2 1 −1 −q/2 2 c e dq n n/2 q n−1 (n−1)/2 n − 1 Γ( 2 )2 Γ( )2 0 2 r Γ( n2 )2n/2 σ2 c use pdf of χ2n (n−1)/2 n − 1 Γ( n−1 )2 2 r Γ( n2 ) 2 ≡σ cσ n − 1 Γ( n−1 ) 2 √ r E(c S 2 ) = = = = = Thus, c = q n−1 n−1 Γ( 2 ) . 2 Γ( n ) 2 5.16 Xi −i 2 i a. P3 b. X1 −1 r P3 Xi −i i=1 i=2 ( i 2 ) ∼ χ23 . ∼ t2 . /2 2 c. rP X1 −1 3 i=2 ( Xi −i 2 /2 i ) ∼ F1,2 . Addtional Prob 1 MXα (t) = 1−t MZα (t) = EetZα α , Xα −EXα t√ = Ee V ar(Xα ) α −α t X√ = Ee α √ √ = e−t α MXα (t/ α) α √ 1 −t α √ = e 1 − t/ α √ √ lnMXα (t) = −t α − αln 1 − t/ α √ t t t2 = −t α − α − √ − + R( √ ) α 2α α 2 t t + αR( √ ) = 2 α 2 t as α → ∞. → 2 3