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

Lecture Notes For Class5

This document discusses continuous random variables. It begins with an example problem calculating probabilities for batches of components with different numbers of defects. It then defines continuous random variables and introduces the concept of a probability density function to assign "weights" instead of probabilities. Examples are given of uniform and non-uniform density functions. Expected value and variance are defined using integrals of the random variable and density function. Homework problems are assigned and example solutions provided.

Uploaded by

Perry01
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
210 views

Lecture Notes For Class5

This document discusses continuous random variables. It begins with an example problem calculating probabilities for batches of components with different numbers of defects. It then defines continuous random variables and introduces the concept of a probability density function to assign "weights" instead of probabilities. Examples are given of uniform and non-uniform density functions. Expected value and variance are defined using integrals of the random variable and density function. Homework problems are assigned and example solutions provided.

Uploaded by

Perry01
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

Continuous Random Variables

May 29, 2014


Continuous Random Variables
Before we begin todays lecture, let us discuss an exercise from the
HW:
Components of a certain type are shipped to a supplier in batches
of ten. Suppose that 50% of all such batches contain no defective
components, 30% contain one defective component, and 20% con-
tain two defective components. Two components from a batch are
randomly selected and tested.
What are the probabilities associated with 0, 1, and 2 defective
components being in the batch under each of the following condi-
tions?
a. Neither tested component is defective.
b. One of the two tested components is defective.
Continuous Random Variables
Let B
i
, i 0, 1, 2 be the event that the batch has i defective
components (out of ten).
We know that PpB
0
q .5, PpB
1
q .3, PpB
2
q .2
Let S
i
, i 0, 1, 2 be the event that the sample has discovered i
defective components (out of two).
We will work only part a) -part b) is similar -. What is being asked
in part a) is
PpB
i
|S
0
q ??
As usual, it helps to visualize the partition of the sample space
with the tree of conditional events.
Continuous Random Variables
Figure: The tree of Conditional Events need pruning.
Continuous Random Variables
Figure: The tree, pruned. We need to nd out the unknowns.
Continuous Random Variables
The unknowns are the conditional probabilities PpS
i
|B
j
q. We need
to compute all of them:
For PpS
0
|B
0
q, it is obvious that if there are 0 defective compo-
nents in the batch, there can not be any defective components in
the sample, so it is 1.
For PpS
1
|B
1
q, if there is one defective component the probability
of detecting it in a sample of two is:
9
`
10
2
.2 PpS
1
|B
1
q
Hence, PpS
0
|B
1
q .8.
Continuous Random Variables
For PpS
2
|B
2
q, we have that if there are two defective components
in the Batch, the probability of detecting both in a sample of two
is
PpS
2
|B
2
q
2
10

1
9
.022
(Think of two independent events of nding one defective compo-
nent at a time) similarly,
PpS
0
|B
2
q
8
10

7
9
.622
Hence, PpS
1
|B
2
q 1 .022 .622 .356
We have found all the unknowns in the tree so we can apply Bayes
theorem as usual.
Continuous Random Variables
PpB
0
|S
0
q
PpB
0
XS
0
q
PpS
0
q

PpS
0
|B
0
q PpB
0
q
PpS
0
XB
0
q `PpS
0
XB
1
q `PpS
0
|B
2
q
all this data is by now known and easy to nd in the tree.
We obtain:
PpB
0
|S
0
q
.5
.5 `.24 `.1244
.578
and, by an analogous reasoning:
PpB
1
|S
0
q
.24
.5 `.24 `.1244
.278
PpB
2
|S
0
q
.1244
.5 `.24 `.1244
.144
Continuous Random Variables
A random variable is continuous if both of the following apply:
1. Its set of possible values consists either of all numbers in a sin-
gle interval on the number line, or a disjoint union of intervals.
2. No possible value of the variable has positive probability, that
is, PpX cq 0 for any value c.
Continuous Random Variables
Figure: Example of a continuous r.v Average temperature on the surface
of the earth.
Continuous Random Variables
Problem: In order to obtain Expected Value, Variance etc of an
r.v. we need the weights ppxq.
But, in a continuous r.v. ppxq 0 for all x.
How to solve this? We introduce the concept of probability density
function.
Let X be a continuous rv. Then a probability distribution or prob-
ability density function (pdf) of X is a function f pxq such that for
any two numbers a and b:
Ppa X bq

b
a
f pxq dx
Continuous Random Variables
Example: The current in a certain circuit as measured by an am-
meter is a continuous random variable X with the following den-
sity function:
f pxq
#
.075x `.2 if 3 x 5
0 otherwise
a. Graph the pdf and verify that the total area under the density
curve is indeed 1.
b. Calculate PpX 4q. How does this probability compare to
PpX 4q?
c. Calculate Pp3.5 X 4.5q and also Pp4.5 Xq.
Continuous Random Variables
a.- The pdf is an ane curve. To check that the total mass is in-
deed 1:

5
3
.075x `.2 dx .0375x
2
`.2x

5
3
. . . 1
b.- PpX 4q PpX 4q PpX 4q. Since X is a continuous
r.v. this is zero.
PpX 4q

4
3
.075x `.2 dx . . . .4625
Continuous Random Variables
We have:
f pxq 0 for all x
and

8
8
f pxq dx 1
Constructing pdf of a continuous r.v. for dierent experiments will
be more delicate than in the discrete case. However, there is a
very elementary distribution that we now describe:
Problem: We wish to nd the pdf of a distribution that, given an
interval rA, Bs R, has PprA, Bsq 1, and the probability of any
subinterval depends only in the length of the subinterval.
Continuous Random Variables
Answer: Clearly, one must have f pxq cte. To see exactly which
constant, we recall that PprA, Bsq 1, so
pB Aq cte 1
We call this the uniform distribution supported on rA, Bs. Thus, it
has pdf:
f pxq
#
pB Aq
1
, if x P rA, Bs
0, otherwise
Continuous Random Variables
Example: Suppose the reaction temperature X (in degrees Cel-
sius) in a certain chemical process has a uniform distribution with
A 5 and B 5.
a. Compute PpX 0q.
b. Compute Pp2.5 X 2.5q.
d. For k satisfying 5 k k `4 5, compute
Ppk X k `4q
Continuous Random Variables
Answer: Of course, one can compute the desired probabilities
with the integral of the (constant) density, but it is faster to use
the characterization of the uniform distribution.
Here, the interval where the distribution is supported is r5, 5s, So
Pp5 X 5q 1
Clearly, then,
PpX 0q 1{2 , Pp2.5 X 2.5q 1{2
(since both intervals are of the same length equal to one half of
the length of the supporting interval), and
Ppk X k `4q 0.4
Continuous Random Variables
HW: Section 4.1, 5, 6 ,10
(ans. for 6: b) k 3{4, d) 0.367 , e) .3125)
(ans. for 10: d) p{aq
k
p{bq
k
)
Continuous Random Variables
Question: How to dene the Expected Value and the Variance?
Answer: Recall that ppxq f pxq dx. Hence,

X
ErXs

8
8
x f pxq dx
If hpXq is a function of the r.v. X, then as before we can consider
the r.v. hpXq. Similarly to the discrete case, we have:
ErhpXqs

8
8
hpxq f pxq dx
Continuous Random Variables
As for the Variance:
VrXs E

pX q
2

8
8
px q
2
f pxq dx
As in the discrete case, it measures the average deviation from the
mean
X
of the r.v.
It can be shown, and it is easier to compute:
VrXs ErX
2
s pErXsq
2
Continuous Random Variables
HW: Section 4.2 , 15, 17, 21, 22.
(ans. for 15: b) 0.0107 , e) ErXs .8182, VrXs 0.0124).
(ans for 21: 314.79 m
2
)
(ans for 22: b) median = 1.640 , c) ErXs 1.614, d) 0.061 )
Continuous Random Variables

You might also like