Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Scribd
Upload
59 views

1 Academic Integrity and Collaboration Policy For This Assignment

This document provides instructions for an optional math homework assignment. It is divided into 4 exercises worth a total of 44 points. The first exercise provides background on Chebyshev's inequality and asks students to use it to find probabilities related to random variables. The second exercise asks students to apply properties of variance to random variables. The third exercise provides background on the central limit theorem and asks students to show properties of sample means, proportions, and totals as they relate to the normal distribution. The fourth exercise asks students to further apply the central limit theorem to random variables. Completing any part of the assignment can increase a student's homework average for the semester. Human: You are an expert at summarizing documents. You provide concise

Uploaded by

Funmath
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
59 views

1 Academic Integrity and Collaboration Policy For This Assignment

This document provides instructions for an optional math homework assignment. It is divided into 4 exercises worth a total of 44 points. The first exercise provides background on Chebyshev's inequality and asks students to use it to find probabilities related to random variables. The second exercise asks students to apply properties of variance to random variables. The third exercise provides background on the central limit theorem and asks students to show properties of sample means, proportions, and totals as they relate to the normal distribution. The fourth exercise asks students to further apply the central limit theorem to random variables. Completing any part of the assignment can increase a student's homework average for the semester. Human: You are an expert at summarizing documents. You provide concise

Uploaded by

Funmath
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

Math 501 Assignment HW EC (optional, unless you wish to try to

increase your semester average): due Thursday 12/19 at 11:59 p.m.


submitted via Canvas in pdf format

Scoring: This assignment is worth 44 points (8 + 4 + 20 + 12 = 44). Your


point score will be added to your homework total for the semester. Thus, even
solving one part of one of the questions can increase your homework average
(even if you do not solve any others).

1 Academic integrity and collaboration policy


for this assignment
You may work together on course material relevant to these problems, and
you are encouraged to do so, but you must complete the written assignment
entirely on your own. Prove you have learned the material well with the written
assignment. Solve these entirely on your own (do the best you can). Neither
our TAs or I will give hints on these. If you do not understand what a question
is asking, let me know, and I will clarify the question.
If you …nd solutions to one or more of the written exercises in Weiss’s book,
in Hsu’s book, or in documents I myself have previously posted to Canvas for
our course this semester, you may make use of them, but the write-up should
be your own, and you must cite the source.
If you …nd solutions to one or more of the written exercises in any source
other than Weiss’s book, Hsu’s book, or documents I myself have previously
posted to Canvas for our course this semester, you may not make use of them
at all, and doing so would be an academic integrity violation.

2 Reading assignment
none

3 Written Assignment (to be turned in)


3.1 Exercise 1 (8 points –4 points per part) (Chebyshev’s
…rst inequality)
3.1.1 Background
When X is a non-negative random variable having …nite expectation, and a is
any positive real number, Chebyshev’s …rst inequality tells us that

E (X)
P (fX ag) :
a

1
We can also use the complementary form of this inequality to give a lower bound
for P (fX < ag) :
If we know an nth moment of X (with n > 1) instead of E (X) ; we can still
make use of Chebyshev’s …rst inequality, as follows.

Example 1 Suppose X is a non-negative random variable having a third mo-


ment, and suppose E X 3 = 40: What, if anything, can be said about P (fX 5g)?
What about P (fX < 5g)?

Solution: Let Y = X 3 : Then Y is a non-negative random variable having a


…rst moment, with E (Y ) = E X 3 = 40; so we can apply Chebyshev’s …rst
inequality to Y :

E (Y ) 40
P (fX 5g) = P (fY 125g) = ;
125 125
40 85
P (fX < 5g) 1 = :
125 125
Thus, P (fX 5g) 40=125; and P (fX < 5g) 85=125:

3.1.2 Exercises
a. Suppose X is a random variable having …nite expectation, and suppose
E (X) = 4: What, if anything, can be said about P (fX 10g)? What about
P (fX < 10g)?

b. Suppose X is a non-negative random variable having a fourth moment, and


suppose E X 4 = 25: What, if anything, can be said about P (fX 3g)? What
about P (fX < 3g)?

3.2 Exercise 2 (4 points – 4 points per part) (Variance


inequality)
3.2.1 Background
The variance of any random variable with a second moment is always non-
negative (which is why the standard deviation, which is the square root of
variance, can be de…ned).
I.e., if X has a second moment, then V AR (X) 0: This follows at once from
2 2
the de…nition of variance, since V AR (X) = E (X X) ; and (X X)
is a non-negative random variable (we cannot average non-negative quantities
and get something negative).
2
Variance is also given by the shortcut formula, V AR (X) = E X 2 E (X) :
Since variance is non-negative, we can conclude that
2
E X2 E (X) 0;

2
and so
2
E X2 E (X) :
This holds for any random variable X having a second moment.
If X has a fourth moment, then we can apply the result with Y = X 2 as
follows:
2
h i2
2 2 4
E X4 = E Y 2 E (Y ) = E X 2 E (X) = E (X) ;

4
and in this way we deduce that E X 4 E (X) for any random variable X
having a fourth moment.
We can continue in this fashion with higher powers as well.

3.2.2 Exercise
a. Suppose X is a random variable having an eighth moment, and suppose
E X 4 = 20: What, if anything, can we say about E X 2 ? What about E (X)?
What about E X 8 ?

3.3 Exercise 3 (20 points – 4 points per part) (Central


limit theorem)
3.3.1 Background
Let X be a random variable which takes the value 1 with probability p 2 (0; 1) ;
and which takes the value 0 with probability 1 p: I.e., X Bernoulli (p) :
Then, as we have calculated in the past (see the random variables formula
sheet and previous HW; or simply repeat the calculations, since they are trivial),
we have

= E (X) = p;
2
= V AR (X) = p (1 p) :

Let X1 ; X2 ; : : : ; Xn be a random sample selected from the distribution of X:


Recall that this means that the Xi s are independent and identically distributed
random variables, each having the same distribution as X (the same cdf as X;
and hence the same mean and variance as X; and so on).
Provided n is large enough, the Central Limit Theorem for means guarantees
that p 2
Xn N ; = n ;

at least approximately (and this is true, at least approximately, if we take a


random sample without replacement, provided we also require n = N to be suf-
…ciently small, as we have discussed).
Let pbn denote the proportion of successes in our random sample of size n:
It turns out that Tn and X n have the same standardization (shown in the
class notes). Since pbn is just X n (shown in the class notes; and you’ll show it

3
below in part a.), it follows that pbn also has the same standardization. We’ll let
Wn denote this common standardization (as in the class notes).
In the class notes, we saw how one central limit theorem became two theo-
rems –one about totals, and one about means. In this exercise, we see how we
get a third theorem, one about proportions. In the study of mathematical sta-
tistics, you will often consider quantities such as totals, means, or proportions,
and the Central Limit Theorem will be indispensable.

3.3.2 Exercise
a. Show that pbn = X n : Then …nd E (b pn ) in terms of p; and …nd V AR (b pn ) in
terms of p:
Hint: Start by showing that pbn = X n : Then …nd E (b pn ) in terms of p; and …nd
V AR (bpn ) in terms of p; and use the result above concerning the approximate
distribution of X n :

b. Suppose Z N 0; 12 ; and suppose X = aZ + b; where a 6= 0: Proceed as


in HW 6 Exercise 5f to show that X N b; a2 :

c. In the Central Limit Theorem, we let Wn be the standardization of Tn : Wn


is also equal to the standardization of X n ; and, since pbn = X n ; it is also equal
to the standardization of pbn :
The Central Limit Theorem shows that, under certain conditions, Wn
N 0; 12 at least approximately (with equality in the limit as n ! 1).
Use the fact that X n = pn Z + ; as well as the result from part b. above,
p 2
to explain why X n N ; ( = n) ; at least approximately, provided n is
large enough.

d. In the Central Limit Theorem, we let Wn be the standardization of Tn : Wn


is also equal to the standardization of X n ; and, since pbn = X n ; it is also equal
to the standardization of pbn :
The Central Limit Theorem shows that, under certain conditions, Wn
N 0; 12 at least approximately p (with equality in the limit as n ! 1).
Use the fact that Tn = ( n) Z + n ; as well as the result from part b.
p 2
above, to explain why Tn N n ; ( n) ; at least approximately, provided
n is large enough.

e. In the Central Limit Theorem, we let Wn be the standardization of Tn : Wn


is also equal to the standardization of X n ; and, since pbn = X n ; it is also equal
to the standardization of pbn :
The Central Limit Theorem shows that, under certain conditions, Wn
N 0; 12 at least approximately (with equality in the limit as n ! 1).
Use your results from part a., and the fact that
pbn E (b
pn )
Wn = ;
ST DDEV (bpn )

4
to write (using elementary algebra) pbn in the form pbn = aWn + b; where a and
b will depend on p:
Then use that fact, as well as the result from part b. above, to explain why
2
pbn N ??; (??) ; at least approximately, provided n is large enough. Of
course, …ll in the ?? with appropriate expressions involving p:

3.4 Exercise 4 (12 points – 4 points per part) (Central


limit theorem –normals, normals, everywhere!)
3.4.1 Background
Re-read the background section from HW 8 Exercise 1 very carefully, and re-
member what the Central Limit Theorem tells us about the sum (i.e., total)
of a large number of i.i.d. (independent and identically distributed) random
variables. This exercise gives another way that normal random variables show
up everywhere: they often show up as limits of random variable distributions.

3.4.2 Exercises
a. Suppose Y B (n; p) : From the background section of HW 8 Exercise 1,
we see that Y can be expressed as Y = X1 + X2 + : : : + Xn ; where Xi
Bernoulli (1; p) for each i and the Xi s are independent. Explain why Y is
approximately normal if n is large enough. (Thus, B (n; p) random variables
converge to normal random variables in the limit as n ! 1:)

b. Use the fact (from the background section of HW 8 Exercise 1) that negative
binomial random variables can be written as sums of i.i.d. geometric random
variables (and give details, as I have done in that background section and in part
a. above) to deduce that N B (r; p) random variables become approximately
normal if r is large enough. (Thus, N B (r; p) random variables converge to
normal random variables in the limit as r ! 1:)

c. Use the fact (from the background section of HW 8 Exercise 1) that chi-
square random variables, 2 (v) ; can be written as sums of i.i.d. 2 (1) random
variables (and give details, as I have done in that background section and in
part a. above) to deduce that 2 (v) random variables become approximately
normal if v is large enough. (Thus, 2 (v) random variables converge to normal
random variables in the limit as v ! 1:)

Remark 2 The background section of HW 8 Exercise 1 has other results which


make it possible to state and prove results of the type above for Poisson random
variables (as sums of i.i.d. Poisson random variables), and for Gamma random
variables of the form (n; ) (as sums of i.i.d. Exponential random variables),
using the identical procedure. Study those as well (but you do not need to turn
anything in for them).

You might also like