MLE Practice and Solutions
1 Key Concepts
For the following data, find the sample size n, and then calculate the maximum likelihood estimate of the
relevant parameter. These are the same exact problems from the method of moments practice document.
1. We poll 10 randomly selected people in California and ask whether they like to dance. 8 of them say
yes. What is your estimate of the proportion of people in California that like to dance?
Solution: The sample size here is n = 10. The responses for this survey are either yes or no,
which we can denote as 1 and 0, respectively, which mean this is a Bernoulli random variable and
we would like to estimate the proportion p of all Californians who like to dance. We found that
8
the MLE for p, p̂M LE = x̄, which is the same as the sample proportion. Thus, p̂M LE = 10 = 0.8.
We could visualize the 10 responses as follows: x1 = 0, x2 = 1, x3 = 0, x4 = 1, x5 = 1, x6 = 1, x7 =
1, x8 = 1, x9 = 1, x10 = 0. From this, we can see that the sample proportion is equivalent to
sum10
i=1 xi 8
x̄ = 10 = 10 = 0.8. The MLE solution is the same as the MOM solution.
2. Five people are stranded on a stretch of highway that is an unknown length. The highway starts at
mile-marker 0, and we know that someone is equally likely to be stranded at any point. The five people
are located at mile markers 1, 3, 4, 6, and 8. How long do we think this highway is?
Solution: Our sample size is n=5, and we are given that x1 = 1, x2 = 3, x3 = 4, x4 = 6, x5 = 8.
The problem tells us that all distances between mile marker zero and the unknown length of the
highway are equally likely, so each Xi follows a uniform distribution with parameter θ. We found
in this section that the MLE for θ is the sample maximum x(n) . Thus, θ̂M LE = 8. Based on our
observed locations of the 5 stranded people, we think this highway is approximately 8 miles long.
3. We ask 4 people in France how much cheese they eat in a year. We know cheese consumption in France
is normally distributed, but we do not know it’s average. The four people tell us that they personally
eat 1kg, 2kg, 3kg, and 4kg of cheese per year. What do you think the average cheese consumption in
France is?
Solution: The sample size is n=4, and our observed data is x1 = 1, x2 = 2, x3 = 3andx4 = 4. We
are told cheese consumption follows a normal distribution, and we want to find the MLE of the
parameter µ. We found in this section that µ̂M LE = x̄ = 1+2+3+4
4 = 2.5. Thus, our best guess of
the average cheese consumption in France, based on the MLE is 2.5kg per year, which is the same
as our MOM solution.
1
2 Skills Practice
Suppose we have a random sample X1 , X2 , . . . , Xn from an exponential distribution with parameter λ. The
probability density function of Xi is given by:
f (xi ) = λe−λxi for values of xi ≥ 0.
The joint probability density function of the sample is given by:
n
Y
L(λ) = f (xi )
i=1
Yn
= λe−λxi
i=1
n −λ n
P
=λ e i=1 xi
Taking the natural logarithm of L(λ), we obtain:
n
X
ln L(λ) = n ln λ − λ xi
i=1
To find the maximum likelihood estimator of λ, we differentiate ℓ(λ) = ln L(λ) with respect to λ and set it
equal to zero:
n
d n X
ln L(λ; x1 , x2 , . . . , xn ) = − xi = 0
dλ λ i=1
n
⇒ λ̂MLE = Pn
i=1 xi
Pn
Therefore, the MLE of λ is λ̂MLE = n/ i=1 xi , which is also equal to 1/X̄, where X̄ is the sample mean.
Page 2
