Scribd
Upload
20 views4 pages

Solutions - Week 7

Uploaded by

pranav.garg1006
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
20 views4 pages

Solutions - Week 7

Uploaded by

pranav.garg1006
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Tutorial Solutions :7

Topics : Central Limit Theorem

1. (1) Pr ⁡(X ≤ 1.5) is given by pnorm(1.5,1,sqrt(.25)). You can get 0.8413447 from R. Note
that we use sqrt(.25) because in R, prnorm() accepts the third value as standard
deviation.
( X−1 ) (1.5−1 )
(2) Give that we know =Z . When X =1.5, Z= =1.0. That is, X =1.5,
√0.25 √ 0.25
metaphorically speaking, has the same probability as Z=1.0
(3) Pr ⁡(Z ≤ 1) is given by pnorm(1,0,sqrt(1)), which is 0.8413447. This, as expected, is
the same as the answer of (1).

The main take-way here is that we can use 2 (a normal distribution) to


X N (μ,σ )
( X−μ )
“build” another random variable . We know this new random variable will have
σ
zero mean and unit variance. I can tell you (as a mathematical fact) that this random
( X−μ )
variable also has a normal distribution. Thus, Z= N ( 0 , 1) .
σ
( x−μ ) ( X−μ )
From the above, we know Pr ( X ≤ x )=Pr ⁡(Z ≤ ) (as Z= , we know when
σ σ
( x−μ )
X ≤ x , it holds Z ≤ ). This explains why the two R commands gives the same
σ
answer.

2. The population from which the sample is drawn is not known to be normally distributed.
By central limit theorem, we know that the sample mean is approximately normally
distributed under larger samples. Using n ≥ 30 as a rule of thumb of large samples, we
got:

( )
4.26−4.18
P ( X ≥ 4.26 ) ≈ P Z ≥ =P ( Z ≥0.67 ),
0.84
√ 50
where the answer from the P ( X ≥ 4.26 ) can be obtained as
1-pnorm(4.26,4.18,0.84/sqrt(50))≈ 0.25; from Q1, we know the same answer can also be
obtained from the standard normal P ( Z ≥ 0.67 ) as 1-pnorm(0.67,0,1)¿0.25.

NOTE 1: You can choose to use either the first command or standard normal distribution

NOTE 2: Importantly, for this and Q3, the “ ≈ (approximately equal to) is meant to
emphasize that the normal distribution under CLM is just an approximation of the
distribution of X —meaning that, under the theorem and supposedly a large sample size,

Tutorial Solutions 7 Page 1


the distribution of X becomes only similar to, but still different from the normal
distribution even when “n ≥ 30” is satisfied. Similarly, this rule does not suggest that CLT
will go sour all of the sudden when n=29.

3. (1)

( )
16−18
P ( X ≥16 )=P Z ≥ =P ( Z ≥−1.85 )
7
√ 42
Either use the R command 1-pnorm(-1.85,0,1)≈ 0.9680. Or use the R command 1-
pnorm(16,18,7/sqrt(42)). Same note as Q1.

( )
16−17.5
(2) P ( X ≤16 )=P Z ≤ =P ( Z ≤−2.03 )
7
√ 90
Use either R command pnorm(-2.03,0,1)≈ 0.0211 or
pnorm(16,17.5,7/sqrt(90))≈ 0.0211. Same caveat as Q1.

4. (1)
> runif(6,0,1)
[no standard answer; you results will differ]
> set.seed(100)
> runif(6,0,1)
0.30776611 0.25767250 0.55232243 0.05638315 0.46854928 0.48377074

Here you can see firsthand the effect of setting a seed value. The first runif command is
executed without setting a specific seed value. Thus, the six numbers are totally
unpredictable. By contrast, the second runif command is under its influence: whoever
runs this runif(6,0,1) after set.seed(100) will get these six numbers.

Note that when a seed value is set, it will “fix” all subsequent commands that generate
random variables (e.g., rnorm(), sample()) in the present R session unless you reset the
seed value.

(2)
> set.seed(50)
> rnorm(3,0,1)
[1] 0.54966989 -0.84160374 0.03299794
> rnorm(3,0,1)
[1] 0.5241497 -1.7276041 -0.2778645

First, notice that while the seed value presets the “randomness” in subsequent
Tutorial Solutions 7 Page 2
commands, it certainly will not “fix” the values from the random variable. This can be
seen by comparing the first three and the next three numbers (they are different).
However, running these three commands will always yield these six numbers.

(3)
> set.seed(50)
> runif(3,0,1)
[1] 0.7087271 0.4376599 0.2000049
> rnorm(3,0,1)
[1] 0.7292185 -1.6985301 0.3754676

Again, comparing the above outputs, we see that the seed value will not make the result of
runif(3,0,1) to be the same as that of rnorm(3,0,1). Comparing the result from rnorm(3,0,1)
and that of the second runif(3,0,1) in part b), we certainly will not expect that those two
commands will yield the same three numbers, even though these two commands are in the
same position after the set.seed command.

But, whoever runs these three commands will always get exactly the same result as shown
above.

5. (1) > for (j in 1:5)


{
j^2
}
[1] 1
[1] 4
[1] 9
[1] 16
[1] 25

(2)
# Think of x as the piggy bank that you keep adding values (i.e., i^2) for 200 times
x<-0
for (i in 1:200){
x=x+i^2 # this reads as “let x equal to (the current value of) x plus i^2
}
x
2686700

But of course, this is not the only way. You can get the same answer (if you are not using
For-loop) by vector multiplication
> x=c(1:200)
> x=x^2
> sum(x)
Tutorial Solutions 7 Page 3
(3)
x<-0
for (i in 100:200){
x=x+i^2
}
x
2358350

Or,
> x=c(100:200)
> x=x^2
> sum(x)

6.

(1) # this “x_bar_3” is used to store the sample mean calculated in each of the 1000
iterations
x_bar_3=rep(0,1000)
set.seed(1)
for (idx in 1:1000) {
x_bar_3 [idx] =mean(rbinom(3,10,0.3))
}
hist(x_bar_3)

(2)
x_bar_3=rep(0,1000)
for (idx in 1:1000) {
set.seed(idx)
x_bar_3 [idx] =mean(rbinom(3,10,0.3))
}
hist(x_bar_3)

Tutorial Solutions 7 Page 4

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy