Statistics 701 HOMEWORK ASSIGNMENTS

          Current HW assignment

Texts: required Peter Bickel and Kjell Doksum, Mathematical Statistics, vol.I, 2nd ed., Pearson Prentice Hall, 2007.
         (recommended) V. Rohatgi and A.K. Saleh, An Introduction to Probability and Statistics, 2nd ed., Wiley.

Problem Set 1. The first assignment consists of 7 problems, due in class Monday, Feb. 9.
         Do the following problems in Rohatgi & Saleh:
                  #2 in Sec 9.5, p.486; #2 in Sec. 9.6, p.489; and #4 in Sec. 10.2, p.499.

         Do problems #4.9.1, 4.9.2, 4.9.4, and 4.9.11, pp. 290-293 in Bickel and Doksum.

Problem Set 2. The second assignment consists of 7 problems, due in class Wednesday, Feb. 18.
         First do the worksheet related to best similar tests (counts as 3 problems).
         Then do the following problems in Bickel & Doksum:   #4.4.6, 4.4.7, 4.4.9, 4.6.1

Problem Set 3. The third assignment consists of 7 problems, due in class Wednesday, Mar. 4.
NOTE that due to the campus closure on March 2, the due-date is extended to
Friday March 6 at 5pm. (You could email or fax or drop off the HW.)

Also note the hint for #5.2.4 below, and the modification of problem (IV) [Part (ii) is now dropped,
and for part (iii) you should try to use the Theorem in Bickel-Doksum on global consistency of
minimum contrast estimators to prove the assertion.]


         (I) (Counts as 2) (i) Solve Bickel-Doksum #5.2.4 parts (a)-(b). Then (ii) compare the large-sample
(asymptotic) variance of the estimator of part (b) with that of the maximum-likelihood estimator of ρ .
Hint: It is not easy to integrate directly to calculate the quadrant probability   P(U>0, V>0)
to be equal to an expression involving the arcsin. But you will find that if you write the quadrant
probability as a univariate integral with integrand involving the standard-normal cdf, then its
derivative with respect to   ρ   is easily integrated analytically and can be shown equal to the
derivative of the epxression you want to equate it to. In part (b), there is a missing term   -1/4
following the average of products of indicators, and to do part (b) you should first estimate that a
negligible-in-probability error is made by replacing each of Xbar and Ybar by their respective
means in the indicator-expressions. Also, in the last part of the problem, you may get the
asymptotic variance either by the delta-method applied to a formula for the MLE or, more
conveniently, by using the Fisher Information.

         (II) Bickel and Doksum #5.2.5.
         (III) Bickel and Doksum #5.3.25, 5.3.27, 5.3.33.
NOTE: the last symbol in the first line of 5.3.33(b) should be sigma-hat squared, not mu-hat squared.
         (IV) Suppose that X1,...,Xn are i.i.d. with Xi = (Yi,Zi) where Y's and Z's are scalar and
Yi-a - b Zi  =  εi,   where the variable   εi   is independent of   Zi, and the variable   εi/σ   has a known
symmetric (i.e. even) density with mean 0 and variance 1.   Here   θ = (a, b, &sigma)   is the unknown parameter vector.
         (i) Show that       ρ(X,θ)   =   &Sigmai=1n    { (Yi-a-bZi)2   +   ( (Yi-a-bZi)2   -   σ2)2 }      is a contrast function.
         (ii) Find the minimum contrast estimators in this problem. NOTE: these estimators cannot be found
in closed form , so you should simply omit this part.

         (iii) Show that the minimum contrast estimators in (ii) are globally (weakly) consistent.

Problem Set 4. The fourth assignment consists of 5 problems, due in class Monday, Mar. 23.
         (A) (Counts as 3) Do the Problems (I)-(III) on the 2x2 Table Handout/Worksheet.

         (B) (Counts as 2) Suppose that U and W are respectively mean-0 random-vectors of dimensions p and q, with
         E(UU') = A ;    (pxp matrix)     ,     E(UW') = B     (pxq matrix) ,         E(WW') = C     (qxq matrix)
and assume that the p+q dimensional random vector concatenating U and W has nonsingular variance-covariance matrix    M   .
                  (i) Prove that the upper-left pxp block of    M-1    is    K   =   (A - B C-1 B')-1    .
                  (ii) Prove that for all p-dimensional non-zero vectors   c   ,     c'K-1c   =   min  b in Rq     E(c'U-b'W)2    .
                  (iii) Use (ii) to conclude that     K-1     is the minimum over pxq matrices   H   of variance-covariance matrices
(in the sense of positive-definite matrix ordering, under which A < B iff B-A is positive-definite) of random p-vectors of
the form     U - HW   .
>br>

Problem Set 5. The fifth assignment consists of 4 problems, due Monday, April 20 in class.
         (A) Do # 5.4.14 p. 360. Recall that the inverse Gaussian density has exponential-family properties
which were established in Stat 700 in Problem 1.6.36 (with solution given in Stat 700 -- see Stat 700 web-page.)
         (B) Suppose that iid data pairs   (Xi,Zi)   for   i=1,...,n   are observed, where   Zi ~ Binom(1,1/2)   and given   Zi   ,
the distribution of   Xi   is   Xi ~ Weibull(γ, &lambda eθZi)   . Suppose that the data are analyzed by estimating   θ   using the
MLE   θ* for the incorrect model which postulates given Zi   that the distribution of   Xi   is   Xi ~ Expon(&lambda eθZi)   .
Note that the parameters   θ   and   γ   and   λ   are all unknown in this problem. The definitions of the Weibull and
Exponential distributions in this problem are as follows: if   T ~ Weibull(a,b),   then   for   t>0,   P(T >t) = exp(-bta)  
and if   V ~ Expon(b),   then   P(V>t) = exp(-bt).
               (a) Show that this misspecified model estimator   θ*   has large-sample limit which is   0   when the true
value of   &theta = 0   and which otherwise (for θ non-zero) exists and has the same sign as θ .
               (b) Find the relative efficiency at   θ=0   of the misspecified model estimator   θ*   of   θ   .

         (C) Do # 6.2.4 and 6.2.5, p. 396.

Problem Set 6. The sixth assignment consists of 5 problems, due Wednesday, May 6 in class:
                       #6.1.4, p.423;    #6.3.1, 6.3.5, pp. 428-430;    #6.4.6, p.432;    #6.5.1, p. 4334.

NOTE added 4/30/09: In Problem #6.1.4, there is a typo in the definition of the weights    aj   :
the numerator of the ratio defining    aj    should be the sum from  i=0   to   i=n-j   of   (-c)i (1/(1+c)) (1 - (-c)i+j).
The denominator of the   aj   expression is exactly as given on p.423 of the Bickel and Doksum book.
To do this problem #6.1.4, I suggest that you do part (b) first: the hint is first to find linear combinations
b1i Y1 + b2i Y2 +... + bii Yi   of the   Yk   observations,   k=1,...,i,   which are strictly independent for
different   i ,   that is, which involve only   ei.

NOTE added 5/4/09: In Problem #6.5.1, you should use a general form of the GLM specificatiion, in which:
g(μ) = Zβ   , with   g   1-to-1 and differentiably invertible, with the link possibly noncanonical, and with the
gradient of   A(η)   one-to-one and differrentiably invertible. This is actually not a difficult problem, if you
write the gradient of   η   with respect to   β   as a   nxp   matrix   M  , and if you try to write the Information
and gradient of likelihood as far as possible in terms of   M.