HW2solcor - Chapter 3 11 The densities are p x = exp ηx...

Info iconThis preview shows pages 1–4. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Chapter 3. 11. The densities are p η ( x ) = exp( ηx- A ( η )) sin( x )1 (0 ,π ) ( x ) , where A ( η ) = log Z R exp( ηx ) sin( x )1 (0 ,π ) ( x ) dx. By integration by parts, Z π exp( ηx ) sin( x ) dx = 1 η exp( ηx ) sin( x ) π- 1 η 2 exp( ηx ) cos( x ) π- 1 η 2 Z π exp( ηx ) sin( x ) dx = 1 + exp( ηπ ) η 2- 1 η 2 Z π exp( ηx ) sin( x ) dx, so that we obtain Z π exp( ηx ) sin( x ) dx = 1 + exp( ηπ ) 1 + η 2 > for any η ∈ R . Therefore, the natural parameter space Θ is R , and the densities are p η ( x ) = 1 + η 2 1 + exp( ηπ ) exp( ηx ) sin( x )1 (0 ,π ) ( x ) . 15. From results about a radius of convergence, we have A ( η ) = ln ∞ X x =1 e- ηx x 2 < ∞ for η > . Thus, Θ = R + . For θ ∈ Θ, by Theorem 3.1, ∞ X x =1 e- ηx x 2 = ∞ X x =1 ∂ 2 ∂η 2 e- ηx = ∂ 2 ∂η 2 ∞ X x =1 e- ηx = ∂ 2 ∂η 2 e- η 1- e- η = ∂ 2 ∂η 2 1 e η- 1 = ∂ ∂η- e η ( e η- 1) 2 = e η ( e η + 1) ( e η- 1) 3 . 1 Therefore, we obtain A ( η ) = ln e η ( e η + 1) ( e η- 1) 3 = η + ln( e η + 1)- 3 ln( e η- 1) and p η ( x ) = ( e η- 1) 3 e η ( e η + 1) e- ηx x 2 . From the argument about the cumulant generating function, we have E [ X ] = E [- T ( X )] =- ∂A ( η ) ∂η =- 1- e η e η + 1 + 3 e η e η- 1 = e 2 η + 4 e η + 1 e 2 η- 1 Var( X ) = ∂ 2 A ( η ) ∂η 2 = ∂ ∂η 1 + 1 1 + e- η- 3 1- e- η = e- η (1 + e- η ) 2 + 3 e- η (1- e- η ) 2 = 4 e- η ( e- 2 η + e- η + 1) ( e- 2 η- 1) 2 19. a) Suppose that | h | < K for some K > 0. Also, since h is differentiable, for a > 0, there exists a δ > 0 such that | h ( x ) | x = | h ( x )- h (0) | x < | h (0) | + for any 0 < x < δ. 2 From the above inequalities, we obtain that Z ∞ | h (1 /x 2 ) | dx = Z ∞ 1 / √ δ + Z 1 / √ δ ! | h (1 /x 2 ) | dx < Z ∞ 1 / √ δ ( | h (0) | + ) 1 x 2 dx + Z 1 / √ δ Kdx = √ δ ( | h (0) | + ) + K 1 √ δ < ∞ . b) By DCT which is justified by a), lim n →∞ n E h (1 / ( n 2 Z 2 ) = lim n →∞ Z ∞-∞ n √ 2 πn 2 h (1 /x 2 ) exp- x 2 2 n 2 dx = r 2 π lim n →∞ Z ∞ h (1 /x 2 ) exp- x 2 2 n 2 dx = r 2 π Z ∞ h (1 /x 2 ) lim n →∞ exp- x 2 2 n 2 dx = r 2 π Z ∞ h (1 /x 2 ) dx. 20. a) f ( x ) = lim n →∞ f n ( x ) = 0. b) Let g = ∑ ∞ n =1 | f n | , which satisfies | f n | ≤ g,n = 1 , 2 ,... . Suppose ∃ g * ( m ) < g ( m ) = c m for some point m ∈ N . Then, g * doesn’t satisfy | f n | ≤ g * ,n = 1 , 2 ,... , in particular g * ( m ) < | f m ( m ) | = c m . Thus, g is the smallest function bounding...
View Full Document

This note was uploaded on 04/14/2010 for the course STATS 610 taught by Professor Moulib during the Fall '09 term at University of Michigan.

Page1 / 11

HW2solcor - Chapter 3 11 The densities are p x = exp ηx...

This preview shows document pages 1 - 4. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online