This is a very elegant sample problem from ISI MStat PSB 2010 Problem 10. It's mostly based on propertes of uniform, and its behaviour when modified . Try it!
Let be a random variable uniformly distributed over
,
, and
.
(a) Find .
(b) Let be a random sample from the above distribution with unknown
. Find two distinct unbiased estimators of
, as defined in (a), based on the entire sample.
Uniform Distribution
Law of Total Expectation
Unbiased Estimators
Well, this is a very straight forward problem, where we just need to be aware the way is defined.
As, we need and by definition of
, we clearly see that
is dependent in
where
.
So, using Law of Total Expectation,
, why ??
Also, conditional pdf of is,
. [where
is the pdf of
].
the other conditional pdf is also same due to symmetry.(Verify!!).
So, .
hence, .
Now, for the next part, one trivial unbiased estimator of is
(based on the given sample). So,
is an obvious unbiased estimator of
.
For another we need to change our way of looking on conventional way and look for the order statistics, since we know that is sufficient for
.(Don't Know ?? Look for Factorization Theorem .)
So, verify that .
Hence, is another unbiased estimator of
. So,
is also another unbiased estimator of
a defined in (a).
Hence the solution concludes.
Let us think about some unpopular but very beautiful relationship between discrete random variables besides the Universality of uniform. Let be a discrete random variable with cdf
and define the random variable
.
Can you verify that, is stochastically greater that a uniform(0,1) random variable
. i.e.
for all
,
,
, for some
,
.
Hint: Draw a typical picture of a discrete cdf, and observe the jump points ! you may jump to the solution!! Think it over.
This is a very elegant sample problem from ISI MStat PSB 2010 Problem 10. It's mostly based on propertes of uniform, and its behaviour when modified . Try it!
Let be a random variable uniformly distributed over
,
, and
.
(a) Find .
(b) Let be a random sample from the above distribution with unknown
. Find two distinct unbiased estimators of
, as defined in (a), based on the entire sample.
Uniform Distribution
Law of Total Expectation
Unbiased Estimators
Well, this is a very straight forward problem, where we just need to be aware the way is defined.
As, we need and by definition of
, we clearly see that
is dependent in
where
.
So, using Law of Total Expectation,
, why ??
Also, conditional pdf of is,
. [where
is the pdf of
].
the other conditional pdf is also same due to symmetry.(Verify!!).
So, .
hence, .
Now, for the next part, one trivial unbiased estimator of is
(based on the given sample). So,
is an obvious unbiased estimator of
.
For another we need to change our way of looking on conventional way and look for the order statistics, since we know that is sufficient for
.(Don't Know ?? Look for Factorization Theorem .)
So, verify that .
Hence, is another unbiased estimator of
. So,
is also another unbiased estimator of
a defined in (a).
Hence the solution concludes.
Let us think about some unpopular but very beautiful relationship between discrete random variables besides the Universality of uniform. Let be a discrete random variable with cdf
and define the random variable
.
Can you verify that, is stochastically greater that a uniform(0,1) random variable
. i.e.
for all
,
,
, for some
,
.
Hint: Draw a typical picture of a discrete cdf, and observe the jump points ! you may jump to the solution!! Think it over.