$25
Let IlJ>0 and let x l,X2 , X n be a random sample from a probability distribution with probability density function zero otherwise.
x 1
Recall: -In 1 has an Exponential distribution with mean 0
2
is an unbiased estimator of V.
k) Suggest a confidence interval for with ( 1 u) 100 % confidence level.
n
O Use Y = E — In 1
2
If T has a Gamma(u, 0 1//X) distribution, then
2 T
//0 = 2 XT has a 1 2 (20) distribution.
l) Suppose n = 3, and x I = 0.62, x 2 = 1.54, x 3 = 1.86.
Use part (k) to construct a 90% confidence interval for V.
m) Find a sufficient statistic u (X 1 , X2, , X n) for V.
n) Find the Fisher information I (y ).
o) Isan efficient estimator of V ?
Ifis not efficient, find its efficiency.
O Find var( ). ("Hint": Recall Homework #08 problem 7 part (g). )
Find the Rao-Cramér lower bound.
Is RIJ an efficient estimator of V? Does Var( RIJ ) attain the R.C.L.B.?
If y is not efficient, find its efficiency.
8. Let — > 0 and let X 1 , X [1] , X n be a random sample from a probability distribution with probability density function
1 4 11
zero elsewhere.
2
1
Recall: W = X 3 has a Gamma( u = 4, 0 ) distribution.
is an unbiased estimator of .
n
i=l
h) Suggest a confidence interval for with ( I — u) 100 % confidence level.
Use Y = X
If T has a Gamma(a, 0= 1/4) distribution, then
i) Suppose n = 5, x 1 = 0.3, x 2 = 0.6, 1.2, x 4 = 1.3, x 5 = 1.8
Use part (h) to construct a 90% confidence interval for g.
j) Find a sufficient statistic u (X , X2, , X n) for g.
k) Find the Fisher information I ( ).
( After you are done with part (k), glance back at Homework #09 problem 8 part (g). )
l) Is an efficient estimator of < ?
Ifis not efficient, find its efficiency.
O Find var( ). ("Hint": Recall Homework #08 problem 8 part (d). )
Find the Rao-Cramér lower bound.
Is an efficient estimator of F, ? Does Var( ) attain the R.C.L.B.?
Ifis not efficient, find its efficiency.
9. Let X > 0 and let X 1 , X 2 , X n be a random sample from a probability distribution with probability density function
2 ' zero otherwise. x
d) Find a sufficient statistic , x n) for X.
[1] T
//0 = 2 XT has a 1 2 (20) distribution.