Chapter Goals

After completing
Aft l ti this
thi chapter,
h t you should
h ld be b
able to:
Engineering Statistics
• Explain three approaches to assessing
probabilities
• Apply common rules of probability
• Use Bayes’ Theorem for conditional probabilities
Chapter 4
Using Probability and • Distinguish between discrete and continuous
Probability Distributions probability distributions
• Compute
C t th
the expected
t d value
l and d standard
t d d
deviation for a discrete probability distribution
1 of 43 2 of 43

Important Terms Sample Space

The Sample Space is the collection of all
• Probability
y – the chance that an uncertain
possible outcomes.
outcomes
event will occur (always between 0 and 1)
• Experiment – a process of obtaining e g All 6 faces of a die:
e.g.
outcomes for uncertain events
• Elementary Event – the most basic
outcome possible from a simple e.g. All 52 cards of a bridge deck:
experiment
• Sample
S l S
Space – the
th collection
ll ti off allll
possible elementary
p y outcomes
3 of 43 4 of 43
 Events Vi
Visualizing
li i E Events
t
• Contingency Tables
• Elementary event – An outcome from a sample
space with one characteristic Ace Not Ace Total

Black 2 24 26
– Example: A red card from a deck of cards
Red 2 24 26
• Event – May involve two or more outcomes Total 4 48 52
simultaneously • Tree Diagrams
Sample
– Example: An ace that is also red from a deck of 2 Space
Sample
cards Space 24
Full
F ll Deck
D k
of 52 Cards 2

5 of 43
24 6 of 43

El
Elementary
t Events
E t Probability Concepts
• An automobile consultant records fuel type and
yp for a sample
vehicle type p of vehicles • Mutually
y Exclusive Events
2 Fuel types: Gasoline, Diesel – If E1 occurs, then E2 cannot occur
3 Vehicle types: Truck, Car, SUV – E1 and E2 have no common elements
6 possible elementary events: e1 E2 A card cannot be
e1 Gasoline, Truck Car e2 E1
Black and Red at
e2 Gasoline Car
Gasoline, e3 Red
R d the same time.
e3 Gasoline, SUV Black Cards
e4
e4 Diesel, Truck
Car
Cards
e5
e5 Diesel, Car
e6 Diesel, SUV e6

7 of 43 8 of 43
 Probability Concepts I d
Independent
d t vs. Dependent
D d tE Events
t
• Independent Events
• Independent and Dependent Events E1 = heads on one flip of fair coin
E2 = heads on second flip of same coin
– Independent: Occurrence of one does not Result
R lt off secondd fli
flip d
does nott depend
d d on th
the result
lt
influence the probability of of the first flip.
occurrence off the
th other
th
• Dependent Events
E1 = rain forecasted on the news
– Dependent: Occurrence of one affects the E2 = take umbrella to work
probability of the other Probability of the second event is affected by the
occurrence of the first event
9 of 43 10 of 43

A i i P
Assigning Probability
b bilit R l off P
Rules Probability
b bilit
• Classical Probability Assessment Rules for
Number of ways Ei can occur
Possible Values
P(Ei) =
Total number of elementary events
and Sum

• Relative Frequency of Occurrence Individual Values Sum of All Values

Number of times Ei occurs
Relative Freq. of Ei = k
N 0 ≤ P(ei) ≤ 1
• Subjective Probability Assessment For any event ei
∑ P(e ) = 1
i=1
i

where:
An opinion or judgment by a decision maker about k = Number of elementary events
the likelihood of an event in the sample space
ei = ith elementary event
11 of 43 12 of 43
 Addition Rule for Elementary Events Complement Rule
• The probability of an event Ei is equal to the • The complement of an event E is the collection
sum off the
th probabilities
b biliti off the
th elementary
l t of all possible elementary events not contained
events formingg Ei. in event E. The complement of event E is
represented by E.
• That is, if: E
Ei = {e1, e2, e3} • Complement Rule:
then:
P( E ) = 1 − P(E) E
P(Ei) = P(
P(e1) + P(e
P( 2) + P(e
P( 3)
Or,, ( ) + P(( E ) = 1
P(E)
13 of 43 14 of 43

Additi R
Addition Rule
l ffor T
Two E
Events
t Additi R
Addition Rule
l EExample
l
■ Addition Rule: P(Red or Ace) = P(Red) +P(Ace) - P(Red and Ace)
P(E1 or E2) = P(E1) + P(E2) - P(E1 and E2) = 26/52 + 4/52 - 2/52 = 28/52
Don’t count
the two red
Color aces twice!
+ = Type Red Black Total
E1 E2 E1 E2
Ace 2 2 4
Non-Ace 24 24 48
P(E
( 1 or E2) = P(E
( 1) + P(E
( 2) - P(E
( 1 and E2)
Don’t count common
T t l
Total 26 26 52
elements twice!

15 of 43 16 of 43
 Addition Rule for
y Exclusive Events
Mutually Conditional Probability
• Conditional probability for any two events
• If E1 and E2 are mutually exclusive
exclusive, then
E1 , E2:
P(E1 and E2) = 0 E1 E2
P(E1 and E 2 )
P(E1 | E 2 ) =
So P(E 2 )
P(E1 or E2) = P(E1) + P(E2) - P(E1 and E2)
where ( 2) > 0
P(E
= P(E1) + P(E2)

17 of 43 18 of 43

C diti
Conditional
lPProbability
b bilit EExample
l C diti
Conditional
lPProbability
b bilit EExample
l
((continued))
• Of the cars on a used car lot, 70% have air conditioning
• Of the cars on a used car lot, 70% have air (AC) and 40% have a CD player (CD).
conditioning (AC) and 40% have a CD player 20% of the cars have both.
(CD). 20% of the cars have both. CD No CD Total
AC .2 .5 .7
• What is the probability that a car has a CD
player, given that it has AC ? No AC .2 .1 .3
Total .4
4 .6
6 10
1.0
i.e., we want to find P(CD | AC)
P(CD and
d AC) .2
2
P(CD|AC) = = = .2857
P(AC) .7
19 of 43 20 of 43
 Conditional Probability Example
(continued) For Independent Events:
• Given AC, we only consider the top row (70% of the cars).
Of these,
these 20% have a CD player
player. 20% of 70% is about • Conditional probability for independent
28.57%. events E1 , E2:
C
CD No CD
C Total
AC .2 .5 .7 P(E1 | E 2 ) = P(E1 ) where P(E 2 ) > 0

No AC .2 .1 .3
Total .4 .6 1.0 P(E 2 | E1 ) = P(E 2 ) where P(E1 ) > 0

P(CD and AC) .2

P(CD|AC) = = = .2857
P(AC) .7
7
21 of 43 22 of 43

Multiplication Rules Tree Diagram Example

P(E
( 1 and E3) = 0.8 x 0.2 = 0.16
• Multiplication
M lti li ti rule
l ffor ttwo events
t E1 and
d E2:
P(E1 and E 2 ) = P(E1 ) . P(E 2 |E1 )
Car: P(E4|E1) = 0.5 P(E1 and E4) = 0.8 x 0.5 = 0.40
Gasoline
P(E1) = 0.8
P(E1 and E5) = 0.8 x 0.3 = 0.24

Note: If E1 and E2 are independent,

p , then P(E 2 | E1 ) = P(E 2 ) P(E2 and E3) = 0.2
0 2 x 0.6
0 6 = 0.12
0 12
and the multiplication rule simplifies to Diesel
P(E2) = 0.2 Car: P(E4|E2) = 0.1
P(E2 and E4) = 0.2 x 0.1 = 0.02
P(E d E 2 ) = P(E
( 1 and ( 1 ) . P(E
( 2) P(E3 and E4) = 0.2 x 0.3 = 0.06

23 of 43 24 of 43
 Addition and M
Multiplication
ltiplication Res
Resume
me
Bayes’ Theorem
A di
diagrammaticti representation
t ti off the
th classification
l ifi ti
of various types of events is as follows:
P(Ei )P(B| Ei )
Two events P(Ei | B) =
E1 and E2
P(E1)P(B| E1) + P(E2 )P(B| E2 ) + K + P(Ek )P(B| Ek )

Mutually Exclusive Events Not Mutually Exclusive

P(E1 E2) = 0 P(E1 + E2) = P(E1) + P(E2) – P(E1 E2) • where:
P(E1 + E2) = P(E1) + P(E2) Ei = ith event of interest of the k possible events
B = new event that might impact P(Ei)
Independent Dependent
Events
E t E1 to
t Ek are mutually
t ll exclusive
l i andd collectively
ll ti l
P(E1 E2) = P(E1) . P(E2) P(E1 E2) = P(E2) . P(E1 | E2) exhaustive
or
P(E1 E2) = P(E1) . P(E2 | E1)
25 of 43 26 of 43

Bayes’ Theorem Example Bayes’ Theorem Example

((continued))

• A drilling
g company
p y has estimated a 40% • Let S = successful well and U = unsuccessful well
chance of striking oil for their new well. • P(S) = .4 , P(U) = .6 (prior probabilities)
• Define the detailed test event as D
• A detailed
d t il d ttestt h
has b
been scheduled
h d l d ffor more • Conditional probabilities:
information. Historically, 60% of successful P(D|S) = .6 P(D|U) = .2
wells have had detailed tests, and 20% of • Revised probabilities
unsuccessful
u success u wells e s have
a e had
ad deta
detailed
ed tests
tests. Prior Conditional Joint Revised
Event
Prob. Prob. Prob. Prob.
• Given that this well has been scheduled for a
S (successful)
( f l) .4
4 .6
6 .4*.6
4* 6 = .24
24 .24/.36
24/ 36 = .67
67
detailed test, what is the probability
U (unsuccessful) .6 .2 .6*.2 = .12 .12/.36 = .33
that the well will be successful?
27 of 43
Sum = .36 28 of 43
 B
Bayes’
’ Th
Theorem E
Example
l IIntroduction
t d ti to t Probability
P b bilit
((continued)) Distributions
• Random Variable
• Given the detailed test, the revised probability
of a successful well has risen to .67 from the – Represents a possible numerical value from
a random event
original estimate of .4
4
Random
Variables
Prior Conditional Joint Revised
Event
Prob
Prob. Prob
Prob. Prob
Prob. Prob
Prob. Discrete Continuous
S (successful) .4 .6 .4*.6 = .24 .24/.36 = .67 Random Variable Random Variable
U (unsuccessful) .6
6 .2
2 .6*.2
6* 2 = .12
12 .12/.36
12/ 36 = .33
33

Sum = .36
36
29 of 43 30 of 43

Di
Discrete
t RRandom
d V
Variables
i bl Di
Discrete
t PProbability
b bilit Di
Distribution
t ib ti
• Can only assume a countable number of values
Experiment: Toss 2 Coins. Let x = # heads.
Examples: 4 possible outcomes
Probability Distribution
– Roll a die twice T T x Value Probability
Let x be the number of times 4 comes up 0 1/4 = .25
(then x could be 0, 1, or 2 times) T H 1 2/4 = .50
2 1/4 = .25
25

– Toss a coin 5 times. H T

Probabilitty
.50
50
L t x be
Let b the
th number b off heads
h d
(then x = 0, 1, 2, 3, 4, or 5) H H .25

31 of 43 0 1 2 x 32 of 43
 Di
Discrete
t PProbability
b bilit Di
Distribution
t ib ti Discrete Random Variable
Summary Measures
• A list of all possible [ xi , P(xi) ] pairs • Expected Value of a discrete distribution
xi = Value
V l off Random
R d V
Variable
i bl (O
(Outcome)
t ) (W i ht d A
(Weighted Average))
P(xi) = Probability Associated with Value
E(x) = Σxi P(xi)
• xi’s are mutually exclusive
(no overlap) – Example: Toss 2 coins
coins, x P(x)
• xi’s are collectively exhaustive x = # of heads, 0 .25
( thi lleft
(nothing ft out)
t) compute expected value off x: 1 .50
• 0 ≤ P(x
( i) ≤ 1 for each xi 2 .25
25
• Σ P(xi) = 1 E(x) = (0 x .25) + (1 x .50) + (2 x .25)
= 1.0

33 of 43 34 of 43

Discrete Random Variable Discrete

Di t RRandom
d V
Variable
i bl
Summary Measures ((continued)) Summary Measures ((continued))

• Standard Deviation of a discrete distribution – Example:

E l Toss
T 2 coins,
i x=#h heads,
d
compute standard deviation (recall E(x) = 1)
σx = ∑ {x − E(x)} P(x) 2

σx = ∑ {x − E(x)} P(x) 2

where:
E(x) = Expected value of the random variable σ x = (0 − 1)2 (.25) + (1 − 1)2 (.50) + (2 − 1)2 (.25) = .50 = .707
x = Values of the random variable
P(x) = Probability of the random variable having Possible number of heads
the value of x = 0,, 1,, or 2

35 of 43 36 of 43
 T
Two Discrete
Di R
Random
d V
Variables
i bl C
Covariance
i
• Expected value of the sum of two discrete • Covariance between two discrete random
random variables: variables:

E(x + y) = E(x) + E(y) σxy = Σ [xi – E(x)][yj – E(y)]P(xiyj)

= Σ x P(x) + Σ y P(y)
where:
(The expected
e pected value
al e of the sum
s m of two
t o random xi = possible
ibl values
l off th
the x di
discrete
t random
d variable
i bl
variables is the sum of the two expected yj = possible values of the y discrete random variable
values)
l ) P(xi ,yyj) = joint probability of the values of xi and yj occurring

37 of 43 38 of 43

Interpreting Covariance Correlation Coefficient

• Covariance between two discrete random • The Correlation Coefficient shows the
strength of the linear association between
variables:
i bl two variables

σxy > 0 x and y tend to move in the same direction σxy

ρ=
σxy < 0 x and y tend to move in opposite directions
σx σy
where:
ρ = correlation coefficient (“rho”)
σxy = 0 x and y do not move closely together
σxy = covariance between x and y
σx = standard deviation of variable x
σy = standard deviation of variable y
39 of 43 40 of 43
 IInterpreting
t ti the
th Ch
Chapter S
Summary
Correlation Coefficient
• Described approaches to assessing
• The Correlation Coefficient always falls probabilities
b biliti
between -1 and +1 • Developed common rules of probability
ρ=0 x and y are not linearly related. • Used Bayes’ Theorem for conditional
The farther ρ is from zero, the stronger the linear probabilities
relationship: • Distinguished between discrete and continuous
ρ = +1 x and y have a p
perfect p
positive linear relationship
p probability distributions
ρ = -1 x and y have a perfect negative linear relationship • Examined discrete probability distributions and
their summary measures.
measures
41 of 43 42 of 43

Thank You

43 of 43