Week 4: Learning

Objectives
At the end of Week 4 you should be able to:

Understand the concept of a random variable
and how it relates to probability problems

Work with probability mass functions

Calculate the expected value, variance and
standard deviation of random variables

Solve probability problems involving the
binomial
probability model
 Random
 A
Variables is a function from the
random variable
sample space S to a set of numbers that
makes sense for the sample space

we will use capital letters to denote a r.v., e.g. X

you can define many r.v. for the same sample
space
 Suppose you roll two dice

sample space consists of 36 outcomes
S={(1,1), (1,2),
(2,1), . . ., (6,6)}

one possible r.v. will represent the sum of the
potential outcomes on the two dice, let’s call the
r.v. X

X will have a different value for each point of the2
 The two-dice
example
Out- X Out- X Out- X Out- X Out- X Out- X
value value value value value value
come come come come come come
(1,1) 2 (2,1) 3 (3,1) 4 (4,1) 5 (5,1) 6 (6,1) 7
(1,2) 3 (2,2) 4 (3,2) 5 (4,2) 6 (5,2) 7 (6,2) 8
(1,3) 4 (2,3) 5 (3,3) 6 (4,3) 7 (5,3) 8 (6,3) 9
(1,4) 5 (2,4) 6 (3,4) 7 (4,4) 8 (5,4) 9 (6,4) 10
(1,5) 6 (2,5) 7 (3,5) 8 (4,5) 9 (5,5) 10 (6,5) 11
(1,6) 7 (2,6) 8 (3,6) 9 (4,6) 10 (5,6) 11 (6,6) 12

X is the random variable

3
 Some more
examples
 Select a random student amongst the
class, and measure his/her height in cm.

S=set of students

r.v. X is ‘height’ which is a function from the
set of students to the set of real numbers that
correspond to heights of students
 I toss a coin 3 times, and measure the
number of heads

S= {HHH, HHT, HTH, HTT, THH, THT, TTH, TTT}

Random variable Y is a function from S to {0, 1,
2, 3}
4
 Types of random

variables
Discrete

A r.v. is discrete if the values it can take are finite
or
countably infinite (link to week 1 slides for these)

The examples with the coin and the dice
correspond to
finite cases

A countably infinite case would be e.g. a r.v.
measuring the number of accidents in a year in
the M1
 Continuous

A r.v. is continuous if the values it can take are
uncountably infinite

e.g. the r.v. measuring the height of students 5
 Probability mass function
(pmf)
 Let X be a discrete random variable
 Given any value a in the set of possible
values for X, what is the probability that X
will have that value a?
 Another way to write this:

P(X=a) : the probability that X will
take the value a, or even simpler
P(a)

6
 Definitions and properties of
pmf
 The probability mass function (pmf) of a
discrete r.v. will give us the value of
P(X=a) for each element a in the set of
possible values of the random variable
X
 For a function to be a probability mass
function of a discrete r.v., 2 conditions
must hold:

P(X=a)0, for all possible a

∑ (P(X=a)) = 1, for all possible a

Both conditions should feel familiar!!!

Also, to find the probability that the r.v takes 7
 Exampl
e
p.m.f of X=Sum
X of two dice  What is the p.m.f. for
(the P(X=a) or P(a) the dice-rolling
r.v.)
2 1/36 example, where the
3 2/36 random variable X is
4 3/36 the sum of the two
5 4/36
6 5/36
dice?
7 6/36
8 5/36  How did we find
9 4/36
10 3/36 these probability
11 2/36 values?
12 1/36  Do they sum up to
1? 8
 A graphical
illustration
0.18

0.16

0.14

0.12
Probability of

0.1

0.08

0.06
X

0.04

0.02

0
1 2 3 4 5 6 7 8 9 10 11 12
Values of X=Sum of two dice

9
 Another
example
 Consider the case of tossing a coin 3
times
 S= {HHH, HHT, HTH, HTT, THH, THT, TTH, TTT}
 Let X be a r.v. that records the number of
heads {0, 1, 2, 3}. Find the p.m.f. of X

- Once you have the p.m.f, find

1 OR 2:

10
 Expectation &
distribution parameters
 Using the probability mass function
you can figure three important
parameters:
 Expected Value

The long-term average value that you would
expect to see after an experiment is
repeated a theoretically infinite number of
times
 Variance

The amount of variability you would need to
expect from one set of results of the
experiment to another
 Standard Deviation
11

 Expected value of a
random variable
 The expected value of a random variable X,
E(X), is the long-run theoretical average
value of X
 You can also see it as the weighted
average of all possible values of X,
weighted by how often we expect each
value to occur over the long- term
, where x is a value of the r.v. X,
E( X )   and
x p(x) the probability of observing that value
 Another way to describe E(X) is as the
mean of X, denoted with the letter 
12
 Methodology &
Example
 To find the expected value of a
random variable:

Multiply the value of X by its
probability

Repeat the step for all values of X

Sum the results
 In the example of rolling the two
dice and the
r.v. X recording the sum of the
dice:
E( X )  2(1/ 36)  3(2 / 36)  4(3 / 36)  5(4 / 36) 
6(5 / 36
8(5 / 36)  9(4 / 36)  13
 The variance of a random
variable
 The variance V(X) of a r.v. is the amount of
variability you would expect in the
results after repeating the experiment a
theoretically infinite number of times
V(X) = E(X2)-2 =E(X2)-[E(X)]2
 In other words, this is the difference
between the expected value of X2 and the
square of the expected value of X
 To find E(X2) you just need to calculate
x2p(x) for all x that are values of the r.v. X
 Always, V(X)0

14
 An
example
 Consider the case where we toss a coin 3
times and the r.v. X counts the number of
heads. Calculate E(X) and V(X)

x xp(x) = 0X(1/8) + 1X(3/8) + 2X(3/8) +
E( X )    all
=
3X(1/8)
3/2
V ( X )  E( X 2 )  (E( X )) 2  ( x 2 p(x))  (E( X )) 2
all x

 [(02  (1/ 8))  (12 (3 / 8))  (22 (3 / 8))  (32  (1/
8))]  (3 / 2)2
3/4
15
 Standard
deviation
 The standard deviation of a random variable
is simply the square root of its variance

we use the letter  to denote the standard
deviation


 Standard deviation allows us to easier
interpret the variation within the
outcomes of the random variable.

It shows us the variability of X in the original
units of X and not in their square, as does
V(X) 16
 Probability
Distributions
 A number of probability distributions have
specific characteristics and can be used to
model experiments when certain
conditions are met

they are distributions that occur in many
probability problems, so people have
studied them and their properties well
 Discrete Uniform, Binomial, Normal,
Bernoulli, Poisson, Gaussian, etc.
 For each distribution there are:

Conditions that must be met

Formulas for pmf, E(X), V(X) 17
 The Binomial
probability
distribution
 Binomial means ‘two names’ and
is associated with situations
involving two outcomes, e.g.
success/failure
 The conditions to have a Binomial
model are:

A fixed number of trials, n

The outcome of each trial is either in one of
two groups, e.g. success or failure
 The probability of success is the same in
each trial, let it be p, and therefore 1-p the
probability of failure
18

 Checking the
conditions
 Do the following examples
satisfy the conditions of the
Binomial model?

You toss a coin 10 times and
count the number of heads

You toss a coin until you get 4
heads
 You have a drawer with 10 red pens, 10
blue and 10 black. You take a pen out
and record its colour (then you do not
put it back in). You repeat the process 5
times, and you are interested in the 19
 The pmf of the
Binomial
 The probability mass function of the Binomial
model for a r.v. X is given by:

P(X=x) is the probability of having
exactly x successes, P(X=x) =C(n,x)px(1-
p)n-x, where:

n is the fixed number of trials

x is the specified number of successes, so n-x
is the number of failures

p is the probability of success in any given
trial, so 1-p
is the probability of failure in any given trial

C(n,x) is our known formula for combinations of
x items chosen from n items (here: x successes
chosen from the total of n trials) 20
 Expected value and
variance
 In the Binomial model, for a r.v. X:

E(X)=np, where:

n is the number of trials, and p the
probability of success in any given trial

V(X)=np(1-p), where:

n is the number of trials, and p the
probability of success in any given trial

The standard deviation np(1 p)
=

21
 Summary of
lecture
 In Week 4 we covered:

Random variables

Probability mass function

Expected value, variance and
standard deviation

The Binomial probability
distribution

22