0% found this document useful (0 votes)
99 views4 pages

Tutorial 10 Solution

This document provides solutions to 7 problems related to statistical inference. Problem 1 shows that a certain expression is an unbiased estimator of a parameter in a Bernoulli population. Problem 2 shows that a given expression is a consistent estimator of another parameter in a Bernoulli population. Problem 3 finds the maximum likelihood estimate of a parameter and shows it is biased. Problem 4 finds the MLE of a parameter based on an independent sample from a given distribution. Problems 5-7 construct 95% confidence intervals for population parameters based on sample data.

Uploaded by

Ujjwal Bansal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
99 views4 pages

Tutorial 10 Solution

This document provides solutions to 7 problems related to statistical inference. Problem 1 shows that a certain expression is an unbiased estimator of a parameter in a Bernoulli population. Problem 2 shows that a given expression is a consistent estimator of another parameter in a Bernoulli population. Problem 3 finds the maximum likelihood estimate of a parameter and shows it is biased. Problem 4 finds the MLE of a parameter based on an independent sample from a given distribution. Problems 5-7 construct 95% confidence intervals for population parameters based on sample data.

Uploaded by

Ujjwal Bansal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Indian Institute of Technology (BHU), Varanasi

Department of Mathematical Sciences


MA-201
Tutorial 10

P P
xi −1)
1. Show that xn(n−1)
i(
is an unbiased estimate of θ, for the sample x1 , x2 , . . . , xn drawn
on X which takes the values 1 or 0 with respective probabilities θ and (1 − θ).
Solution: Since 0 x1 , x2 , . . . , xn is a random sample from Bernoulli population with pa-
rameter θ,
n
X
T = xi ∼ B(n, θ)
i=1

This implies

E(T ) = nθ and Var(T ) = nθ(1 − θ)

Therefore,
P P   
xi ( xi − 1) T (T − 1)
E =E
n(n − 1) n(n − 1)
1
E T 2 − E(T )
  
=
n(n − 1)
1
Var(T ) + {E(T )}2 − E(T )
 
=
n(n − 1)
1
nθ(1 − θ) + n2 θ2 − nθ
 
=
n(n − 1)
nθ2 (n − 1)
= = θ2
n(n − 1)
This implies [ xi ( xi − 1)] /[n(n − 1)] is an unbiased estimator of θ2 .
P P

2. If X1 , X2 , . . . , Xn are random observations on a Bernoulli variate X taking the value 1


with probability p and the value 0 with probability (1 − p), show that
 P 
Σxi xi
1−
n n
is a consistent estimator of p(1 − p).
Solution: Since X1 , X2 , . . . , Xn are iid Bernoulli variates with parameter p,
n
X
T = xi ∼ B(n, p)
i=1

This implies

E(T ) = np and Var(T ) = npq


n
1X T
X̄ = xi =
n i=1 n

1
Therefore,
1 1
E(X̄) = E(T ) = · np = p
n  n
T 1 pq
Var(X̄) = Var = 2 · Var(T ) = → 0 as n → ∞
n n n

Since E(X̄) → p and Var(X̄) → 0, as n → ∞; X̄ is a consistent estimator of p. Also


P P 
xi xi
n
1− n = X̄(1 − X̄), being a polynomial in X̄, is a continuous function of X̄.
Since X̄ is consistent estimator of p, by the invariance property of consistent estimators,
X̄(1 − X̄) is a consistent estimator of p(1 − p).

3. Prove that the maximum likelihood estimate of the parameter α of a population having
density function
2
f (x) = (α − x), 0 < x < α
α2
for a sample of unit size is 2x, x being the sample value. Show also that the estimate is
biased.
Solution: For a random sample of unit size (n = 1), the likelihood function is :
2
L(α) = f (x, α) = (α − x); 0 < x < α
α2
Likelihood equation gives :
d d
log L = [log 2 − 2 log α + log(α − x)] = 0
dα dα
2 1
=⇒ − + =0
α α−x
=⇒ 2(α − x) − α = 0
=⇒ α = 2x.

Hence MLE of α is given by α̂ = 2x


Z α Z α α
4 αx2 x3

4 2
E(α̂) = E(2X) = 2 x · f (x, α)dx = 2 x(α − x)dx = 2 − = α
0 α 0 α 2 3 0 3

Since E(α̂) 6= α, α̂ = 2x is not an unbiased estimate of α.

4. Obtain the maximum likelihood estimate of θ in

f (x, θ) = (1 + θ)xθ , 0 < x < 1

based on an independent sample of size n.

2
Solution: Note that
n n

Y Y
L(x, θ) = f (xi , θ) = (1 + θ)n · xi
i=1 i=1
n
X
=⇒ log L = n log(1 + θ) + θ. log xi
i=1
n
∂ n X
=⇒ log L = + log xi = 0
∂θ 1 + θ i=1
X X
=⇒ n+θ log xi + log xi = 0
i i
−n −n
=⇒ θ̂ = Pn −1= Qn − 1.
i=1 log xi log ( i=1 xi )

5. A random sample of size n = 100 is taken from a population with σ = 5.1. Given that the
sample mean is x̄ = 21.6, construct a 95% confidence interval for the population mean µ.
Solution: Substituting the given values of n, x̄, σ, and z0.025 = 1.96 into the confidence
interval formula, we get
5.1 5.1
21.6 − 1.96 · √ < µ < 21.6 + 1.96 · √
100 100
or 20.6 < µ < 22.6.

6. We know that silk fibers are very tough but in short supply. Engineers are making break-
throughs to create synthetic silk fibers that can improve everything from car bumpers
to bullet-proof vests or to make artificial blood vessels. One research group reports the
summary statistics

n = 18 x̄ = 22.6 s = 15.7

for the toughness (MJ/m3 ) of processed fibers. Construct a 95% confidence interval for
the mean toughness of these fibers. Assume that the population is normal.
Solution: The sample size is n = 18 and t0.025 = 2.110 for n − 1 = 17 degrees of freedom.
15.7 15.7
The 95% confidence formula for µ becomes 22.6 − 2.110 · √ 18
< µ < 22.6 + 2.110 · √ 18
or
3
14.79 < µ < 30.41 MJ/m .

7. A major manufacturer of processed meats monitors the amount of each ingredient. The
weight(lb) of cheese per run is measured on n = 80 occasions.
72.2 67.8 78.0 64.4 76.3 72.3 73.1 71.7 66.2 63.3 85.4 67.4
66.3 76.3 57.7 50.3 77.4 63.1 73.9 67.4 74.7 68.2 87.4 86.4
69.4 58.0 63.3 72.7 73.6 68.8 63.3 63.3 73.0 64.8 73.1 70.9
85.9 74.4 75.9 72.3 84.3 61.8 79.2 64.3 65.4 66.7 77.2 50.0
70.3 90.4 63.9 62.1 68.2 55.1 52.6 68.5 55.2 73.5 53.7 61.7
47.9 72.3 61.1 71.8 83.1 71.2 58.8 61.8 86.8 64.5 52.3 58.3
65.9 80.2 75.1 59.9 62.3 48.8 64.3 75.4
Assume the population is normal, construct a 95% confidence interval for the population
standard deviation σ.
Solution: Since the population is normal with n = 80, we can use χ2 . Note that

3
χ2.975 = 56.309 and χ2.025 = 105.473. and then s = 9.583. Substituting into the formula for
the confidence interval for σ 2 yields

(79)(9.583)2 (79)(9.583)2
< σ2 <
105.473 56.309
or

68.8 < σ 2 < 128.8

and, taking the square root,

8.29 < σ < 11.35

You might also like