SME11e PPT ch08std

Slides by:
Andrew Stephenson
Georgia Gwinnett College
© 2018 Cengage Learning. All Rights Reserved. May not be copied, scanned, or duplicated, in whole or in part, except for use as permitted in
a license distributed with a certain product or service or otherwise on a password-protected website for classroom use.
Chapter 8 Continuous Probability
Distributions (연속확률분포) 8.2
• Probability Density Functions (확률밀도함수)
• Normal Distribution (정규분포)
• Exponential Distribution (지수분포)
• Other Continuous Distributions
Introduction
8.3
• In Chapter 7, we introduced discrete probability distributions that are
employed to calculate the probability associated with discrete random
variables.
• In Section 7-4, we introduced the binomial distribution to determine
the probability that the random variable equals a particular value
(the number of successes). In this way we connected the population
represented by the probability distribution with a sample of nominal
data (success or failure).
• In Chapter 8, we introduce continuous probability distributions, which

are used to calculate the probability associated with an interval variable.
• By doing so, we develop the link between a population and a sample
of interval data.
8-1 Probability density function (확률밀도함수)
8.4
• Unlike a discrete random variable, a continuous random variable is
one that can assume an uncountable number of values: It can assume
any value in an interval on the line or in a collection of intervals.
• We cannot list the possible values
in the interval [0, 1] because
there is an infinite number of them.
0 1
e.g. 𝑋 = 0.0000001, …
• It is not possible to talk about the probability of the random
variable assuming an particular value: 𝑃 𝑋 = 𝑥 = 𝑃 𝑥 = 0
• Point probabilities are zero!!
8.5
• Consequently, we can determine the probability of only a range of

values:
e.g. 𝑃(𝑋 ≤ 𝑥1 ), 𝑃(𝑥1 ≤ 𝑋 ≤ 𝑥2 ), 𝑃(𝑋 ≥ 𝑥2 )
8.6
• To illustrate how this is done, consider the histogram we created for the
ages of ACBL members(Example 3.1).
Fig.8.1 Histogram for Example 3.1
Histogram
50
40
Frequency
30
20
10
0
20 30 40 50 60 70 80 90 100
Age (Years)
8.7
• We found, for example, that the relative frequency of the interval 10-20
was 6/200.
• Using the relative frequency approach, Interval Relative
frequency
we estimate that the probability that
10 < 𝑋 ≤ 20 6/200
a randomly selected ACBL members
20 < 𝑋 ≤ 30 27/200
will be between 10 and 20 years of age
30 < 𝑋 ≤ 40 30/200
is 6/200=.030.
40 < 𝑋 ≤ 50 16/200
• Similarly, the probabilities of other 50 < 𝑋 ≤ 60 40/200
intervals in the histogram. 60 < 𝑋 ≤ 70 36/200
70 < 𝑋 ≤ 80 27/200
80 < 𝑋 ≤ 90 12/200
90 < 𝑋 ≤ 100 6/200
Total 200/200 = 1
8.8
• Notice that the sum of the probabilities equals 1.
• We set the values along the vertical axis so that the area in all the
rectangles together adds to 1. (dividing each relative frequency by
the width of the interval, which is 10.)
• The result is a rectangle over each interval whose area equals the
probability that the random variable will fall into that interval.
• The probability that an ACBL member is between 25 and 45 years of

age is equal to the area between 25 and 45 as shown in Fig. 8.2 :
𝑃 25 < 𝑋 ≤ 45 = 5 .0135 + 10 .0150 + 5 .008
=.0675+.015+.04=.2575
8.9
Interval Relative Figure 8.2 Estimated prob. that an ACBL member
frequency/10 will be between 25 and 45
10 < 𝑋 ≤ 20 .0030
Relative
20 < 𝑋 ≤ 30 .0135 freq./10 .020
.018
30 < 𝑋 ≤ 40 .0150
.015
40 < 𝑋 ≤ 50 .0080 .0135 .0135
50 < 𝑋 ≤ 60 .0200
60 < 𝑋 ≤ 70 .0180 .008
.006
70 < 𝑋 ≤ 80 .0135
.003 .003
80 < 𝑋 ≤ 90 .0060
90 < 𝑋 ≤ 100 .0030
20 30 40 50 60 70 80 90 100
Total 200
= .10 Age (Years)
2000
8.10
• If the histogram is drawn with a large number of small intervals, we can
smooth the edges of the rectangles to produce a smooth curve as
shown in Figure 8.3
• In many cases, it is possible to determine a function 𝑓(𝑥) that
approximates the curve. The function is called a probability density
function.
Fig. 8.3 Density Function for Example 3.1
Cumulative Distribution Function(cdf)
8.11
• Definition 8.1 ((Cumulative) Distribution Function*: cdf) Let 𝑋
denote any random variable. The distribution function of 𝑋, denoted
by 𝐹 𝑥 , is given by
𝐹 𝑥 ≡ 𝑃 𝑋 ≤ 𝑥 𝑓𝑜𝑟 − ∞ < 𝑥 < ∞
• It denotes the probability that the random variable 𝑋 takes on a
value which will be less than or equal to 𝑥.
• The nature of the distribution function associated with a random
variable determines whether the variable is continuous or discrete.
* (Cumulative) distribution function: (누적)분포함수
8.12
• Example: Suppose that 𝑋 has a binomial distribution with 𝑛 = 2 and
1
𝑝 = . Find 𝐹(𝑥).
2
• Solution: The probability function for 𝑋 is given by

1 1
𝑝 𝑥 = 𝐶𝑥2 ( )𝑥 ( )2−𝑥 , 𝑥 = 0, 1, 2
2 2
1 1 1
𝑝 0 = , 𝑝 1 = , 𝑝 2 = . So
4 2 4
• Because the only values of 𝑋 that are assigned to positive

probabilities are 0, 1, and 2,
𝐹 −2 = 𝑃 𝑋 ≤ −2 = 0 and
1 2 3
𝐹 1.5 = 𝑃 𝑋 ≤ 1.5 = 𝑃 𝑋 = 0 + 𝑃 𝑋 = 1 = + =
4 4 4
• In general, 8.13
0, 𝑓𝑜𝑟 𝑥 < 0
1
, 𝑓𝑜𝑟 0 ≤ 𝑥 < 1
4
𝐹 𝑥 =𝑃 𝑋≤𝑥 = 3
4
, 𝑓𝑜𝑟 1 ≤ 𝑥 < 2
1, 𝑓𝑜𝑟 𝑥 ≥ 2
F(x)
• The points between 0 and 1 or between
1 and 2 contribute nothing to the 1
cumulative probability.
3/4
• As a result, the cumulative distribution
2/4
function stays flat between the possible 1/4
values of 𝑋.
0 1 2 x
• It increases in jumps or steps at each of
the possible values of 𝑋.
FYI: Properties of the Distribution Function
8.14
• Theorem 8.1 (Properties of the distribution function) If 𝐹 𝑥 is
a distribution function, then
1. 𝐹 −∞ = lim 𝐹(𝑥) = lim 𝑃(𝑋 ≤ 𝑥) = 0,
𝑥→−∞ 𝑥→−∞
2. 𝐹 ∞ = lim 𝐹(𝑥) = lim 𝑃(𝑋 ≤ 𝑥) = 1,
𝑥→∞ 𝑥→∞
3. 𝐹(𝑥) is a non-decreasing function of 𝑥 (If 𝑥1 𝑎𝑛𝑑 𝑥2 are any

values such that 𝑥1 < 𝑥2 , then 𝐹(𝑥1 ) ≤ 𝐹(𝑥2 )).
Continuous Random Variable (연속확률변수)
8.15
• Definition 8.2 (Continuous Random Variable X):
• Let 𝑋 denote a random variable with distribution function 𝐹(𝑥). 𝑋
is said to be continuous if the distribution function 𝐹(𝑥) is
continuous for −∞ < 𝑥 < ∞.
• For a continuous random variable 𝑋, we must have, for any real
number 𝑥, 𝑃 𝑋 = 𝑥 = 0. [If this were not true, and 𝑃 𝑋 = 𝑥0 =
𝑝0 > 0, then 𝐹(𝑥) would have a discontinuity (jump) of size 𝑝0 at
the point 𝑥0 .]
• Continuous random variable 𝑋 can have uncountably infinite
number of values in a given interval.
• Length, weight, time, temperature, etc.: the time that a train
arrives at a specified stop and the life time of a tire.
8.16
• Definition 8. 3 (Probability density function*) A function f(x) is
called a probability density function (pdf) over the range a ≤ x ≤ b if it
meets the following two requirements (see Theorem 8.2):
1. 𝑓(𝑥) ≥ 0 for all x between a and b.
2. The total area under the curve between a and b is 1.0.
f(x)
area=1
a b x
* Integral calculus (적분) can often be used to determine probabilities (calculate the area
under a curve) and parameters.
Probability Density Function…
8.17
• Definition 8.3-1 (Probability Density Function: pdf) Let 𝐹(𝑥) be
the distribution function for a random variable 𝑋. Then 𝑓 𝑥 , given by
𝑑𝐹 𝑥
𝑓 𝑥 ≡ = 𝐹′(𝑥) ≥ 0
𝑑𝑥
wherever the derivative* exists, called the probability density function

for the random variable 𝑋.
• Theorem 8.2 (Properties of a Density Function 𝑓(𝑥))

• 1. 𝑓 𝑥 ≥ 0, ∀ 𝑥** (since 𝐹(𝑥) is a non-decreasing function.)
∞
• 2. ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 = 1 (since 𝐹 ∞ = 1)
* derivative: 도함수
** ∀ 𝑥: for all x
Probability Density Function …
8.18
• Theorem 8.3 (Probability density function and its probabilities) If the
random variable 𝑋 has the density function 𝑓(𝑥) and 𝑥1 ≤ 𝑥2 , then the
probability that 𝑋 falls in the interval [𝑥1 , 𝑥2 ] is
𝑥
𝑃 𝑥1 ≤ 𝑋 ≤ 𝑥2 = ‫ 𝑥׬‬2 𝑓 𝑥 𝑑𝑥
1
Proof:
2 1 𝑥 𝑥
𝐹 𝑥2 − 𝐹 𝑥1 = ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 − ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥
1 𝑥 𝑥 𝑥
= ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 + ‫ 𝑥׬‬2 𝑓 𝑥 𝑑𝑥 − ‫׬‬−∞
1
𝑓 𝑥 𝑑𝑥
1
𝑥
= ‫ 𝑥׬‬2 𝑓 𝑥 𝑑𝑥 ∎*
1
* The second fundamental theorem of calculus. See theorem 8.4 in FYI.
Probability Density Function…
8.19
• If 𝑋 is a random variable and 𝑎 and 𝑏 are constants such that 𝑥1 < 𝑥2 ,
then 𝑃 𝑋 = 𝑥1 = 0 𝑎𝑛𝑑 𝑃 𝑋 = 𝑥2 = 0, which implies the following
facts:
𝑃(𝑥1 ≤ 𝑋 ≤ 𝑥2 ) = 𝑃(𝑥1 ≤ 𝑋 < 𝑥2 ) = 𝑃(𝑥1 < 𝑋 ≤ 𝑥2 ) = 𝑃(𝑥1 < 𝑋 < 𝑥2 )
Probability Density Function and Its
Distribution Function … 8.20
• Example: Suppose that
0, 𝑓𝑜𝑟 𝑥 < 0
𝐹 𝑥 = ቐ𝑥, 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1
1, 𝑓𝑜𝑟 𝑥 > 1 F(x)
F(𝑥2 ) − 𝐹 𝑥1 = 𝑃(𝑥1 < 𝑋 < 𝑥2 )
Q: Find the probability density function for 𝑥, 1

and graph it. F(𝑥2 )
F(𝑥1 )
Solution: 𝑥1 𝑥2 1 x
f(x)
0, 𝑓𝑜𝑟 𝑥 < 0
𝑓 𝑥 =
𝑑𝐹(𝑥)
= ቐ1, 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1 1 𝑥2
𝑑𝑥
0, 𝑓𝑜𝑟 𝑥 > 1 න 1𝑑𝑥
𝑥1
x
𝑥1 𝑥2 1
• Interpretation:
• 𝑓(𝑥) is undefined at 𝑥 = 0, 𝑥 = 1.
• In general, the distribution function for a continuous random variable
must be continuous, but the density function need not be
everywhere continuous.
• Note that 𝑓 𝑥 is a slope of the cumulative distribution function 𝐹(𝑥)
at point 𝑥.
• cdf cannot be greater than 1 but pdf can be greater than 1
because pdf is the slope of cdf.
FYI: Indefinite (Definite) Integration (부정(정)적분…
8.22
• Definition (indefinite integral) In calculus, an indefinite integral (or
antiderivative) of a function f is a differentiable function F whose
derivative (도함수) is equal to the original function f.
• Symbolically, 𝐹 ′ = 𝑓.
• The process of solving for indefinite integrals is called indefinite
integration, and its opposite operation is called differentiation (미분),
which is the process of finding a derivative.
• Indefinite integrals are related to definite integrals through the second
fundamental theorem of calculus.
• The second fundamental theorem of calculus: the integral of a
function f over a fixed interval is equal to the change of any indefinite
integral F between the ends of the interval.
𝑥 𝑥2
‫ 𝑥׬‬2 𝑓 𝑥 𝑑𝑥 = 𝐹(𝑥) 𝑥1 = 𝐹 𝑥2 − 𝐹(𝑥1 )
1
FYI: Indefinite (Definite) Integration (부정(정)적분…
8.23
Example:
𝑥3 𝑥2
• 𝐹 𝑥 =E − − 𝑥 + 𝑐x
3 2
• Showing three of infinitely many solutions that can be

produced by varying the arbitrary constant c.
• c: the constant of integration (적분상수), c = 4, 0, -4
• Example: Let 𝑋 be a random variable with probability density function
given by f(x)
3𝑥 2 , 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1 3
𝑓 𝑥 =ቊ
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
• Graph of 𝑓(𝑥):
0 1 x
• Find 𝐹 𝑥 𝑎𝑛𝑑 graph 𝐹 𝑥 .
𝑥
• Solution: Because 𝐹 𝑥 = ‫׬‬−∞ 𝑓 𝑡 𝑑𝑡, we have
𝑥
‫׬‬−∞ 0𝑑𝑡 = 0, 𝑓𝑜𝑟 𝑥 < 0
0 𝑥
𝐹 𝑥 = ‫׬‬−∞ 0𝑑𝑡 + ‫׬‬0 3𝑡 2 𝑑𝑡 = 0 + 𝑡 3 ]0𝑥 = 𝑥 3 , 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1
0 𝑥 𝑥
‫׬‬−∞ 0𝑑𝑡 + ‫׬‬0 3𝑡 2 𝑑𝑡 + ‫׬‬1 0𝑑𝑡 = 0 + 𝑡 3 ]10 ) + 0 = 1, 𝑓𝑜𝑟 1 < 𝑥.
F(x)
0 1 x
Distribution Function… 8.26
• Example: Let the distribution function of a random variable 𝑋 be
0, 𝑥<0
𝑥
, 0<𝑥<2
8
𝐹 𝑥 =
𝑥2
, 2≤𝑥≤4
16
1, 𝑥≥4
a. Find the density function of 𝑋.
b. Find 𝑃 1 ≤ 𝑋 ≤ 3 .
c. Find 𝑃 𝑋 ≥ 1.5 .
d. Find 𝑃 𝑋 ≥ 1 𝑋 ≤ 3
Distribution Function… F(x)
8.27
Solution: 1
0, 𝑥<0
1
, 0<𝑥<2 9/16
8
a. 𝑓 𝑥 = 1
𝑥, 2≤𝑥≤4 3/8
8 2/8
0, 𝑥≥4 1/8
1 2 3 4 x
f(x)
1
4/8 3
2/8 න 𝑓(𝑥)𝑑𝑥
1/8 1
1 2 3 4
x
Distribution Function… 8.28
Solution:
0, 𝑥<0
1
, 0<𝑥<2
8
a. 𝑓 𝑥 = 1
𝑥, 2≤𝑥≤4
8
0, 𝑥≥4
3 21 3𝑥
b. 𝑃 1 ≤ 𝑋 ≤ 3 = ‫׬‬1 𝑓 𝑥 𝑑𝑥 = ‫׬‬1 𝑑𝑥 + ‫׬‬2 𝑑𝑥
8 8
𝑥 2 𝑥2 3 1 5 7
= ] + ] = + = ∎
8 1 16 2 8 16 16
1 3 13
c. 𝑃 𝑋 ≥ 1.5 = 1 − 𝑃 𝑋 < 1.5 = 1 − = ∎
8 2 16
7
𝑃(1≤𝑋≤3) (16) 7
d. 𝑃 𝑋 ≥ 1 𝑋 ≤ 3 = = 9 = ∎
𝑃(𝑋≤3) (16) 9
𝑥2 4 7 9
𝑃 𝑋 ≤ 3 = 1 − 𝑃 𝑋 > 3 = 1 − ]3 = 1 − = ∎
16 16 16
Expected Values for Continuous Variables…
8.29
• Definition 8.4 (Expected value of a continuous random variable)
The expected value of a continuous random variable 𝑋 is
∞
𝐸 𝑋 = 𝜇 = ‫׬‬−∞ 𝑥𝑓 𝑥 𝑑𝑥
provided that the integral exists.
• As numerical descriptive measures, we need to find means

and variances of continuous random variables.
• Many times it is difficult to find the probability distributions for
a random variable 𝑋 or a function of a random variable 𝑔(𝑋).
• Even if the density function is known , it can be difficult to
evaluate appropriate integrals (e.g. gamma distribution).
Expected Value of a Function of a
Random Variable… 8.30
• Theorem 8.4 (The Expected value of a function of a random
variable) Let g(𝑋) be a function of 𝑋. Then the expected value of
𝑔(𝑋) is given by
∞
𝐸[𝑔 𝑋 ] = න 𝑔(𝑋)𝑓 𝑥 𝑑𝑥
−∞
provided that the integral exists.
Expected Value of a Function of a
Random Variable… 8.31
• Theorem 8.5 (The Expected values of three important functions)
Let g(𝑋) and 𝑔1 𝑋 , 𝑔2 𝑋 , … , 𝑔𝑘 (𝑋) be a functions of a continuous
random variable 𝑋. Then the following results hold:
1. 𝐸 𝑐 = 𝑐. (𝑐: a constant)
2. 𝐸 𝑐𝑔 𝑋 = 𝑐𝐸[𝑔 𝑋 ].
3. 𝐸 𝑔1 𝑋 + 𝑔2 𝑋 + ⋯ + 𝑔𝑘 𝑋 = 𝐸 𝑔1 𝑋 + 𝐸 𝑔2 𝑋 + ⋯ + 𝐸[𝑔𝑘 𝑋 ]
• The expected value of function 𝑔 𝑋 = (𝑋 − 𝜇)2 is the

variance of 𝑋, 𝑉(𝑋).
2 ∞
𝑉 𝑋 =𝐸 𝑋−𝜇 = ‫׬‬−∞ 𝑋 − 𝜇 2 𝑓 𝑥 𝑑𝑥
Expected Value… 8.32
• Example: Given 𝑓 𝑥 = 𝑐𝑥 2 , 0 ≤ 𝑥 ≤ 2, 𝑎𝑛𝑑 𝑓 𝑥 = 0 elsewhere.
1) Find the value of 𝑐 for which 𝑓 𝑥 is a valid density function.
2) Find 𝑃(1 ≤ 𝑋 ≤ 2)
3) Find 𝜇 = 𝐸 𝑋 𝑎𝑛𝑑 𝜎 2 = 𝑉 𝑋
Solution:
Expected Value… 8.33
Solution:
1) We require a value for 𝑐 such that
∞ 2 𝑐𝑥 3 2 8𝑐 3
𝐹 ∞ = ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 = 1 = ‫׬‬0 𝑐𝑥 2 𝑑𝑥 =
3 0
] =
3
−0 ⇒𝑐 = ∎
8
23 3 𝑥3 2 1 7
2) 𝑃 1 ≤ 𝑋 ≤ 2 = ‫׬‬1 8 𝑥 2 𝑑𝑥 = ] = 8−1 = ∎
8 3 1 8 8
2 2 3 2 3 2 3 3 𝑥4 2 3
3) 𝐸 𝑋 = ‫׬‬0 𝑥𝑓 𝑥 𝑑𝑥 = ‫׬‬0 𝑥 𝑥 𝑑𝑥 = ‫𝑥𝑑 𝑥 ׬‬ = ] = ∎
8 8 0 8 4 0 2
2 2 2 3 2 4 3 𝑥5 2 12
𝐸 𝑋 = ‫׬‬0 𝑥 𝑓 𝑥 𝑑𝑥 = ‫ =𝑥𝑑 𝑥 ׬‬8 5 ]0 = ,
8 0 5
12 9 3
𝑉 𝑋 = 𝐸 𝑋2 − 𝐸 𝑋 2 = − = ∎
5 4 20
Various Continuous Random Variables…
8.34
• Uniform Distribution
• Normal Distribution
• Exponential Distribution
• Chi-squared Distribution
• t-Distribution
• F-Distribution
• etc.
8-1a Uniform Distribution (균일분포, 일양분포) …
8.35
• Definition 8.5 (Uniform Distribution) A random variable 𝑋 is said
to be uniformly distributed on the interval(𝑎, 𝑏) if its density function
is given by
1
, 𝑎≤𝑥≤𝑏
𝑓 𝑥 = ൝𝑏−𝑎
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒
⇒ whenever the probability is proportional to the interval’s length.
• Notation: 𝑋~𝑈(𝑎, 𝑏) or 𝑋~𝑈[𝑎, 𝑏]
𝑥
• Distribution function: 𝐹 𝑥 =
𝑏−𝑎
0, 𝑘≤𝑎
𝑘 𝑘−𝑎
𝐹 𝑘 = ‫ = 𝑥𝑑 𝑥 𝑓 𝑎׬‬൞𝑏−𝑎 , 𝑎 < 𝑘 ≤ 𝑏
1, 𝑘≥𝑏
8-1a Uniform Distribution (균일분포, 일양분포) …
8.36
• To calculate the probability of any interval, simply find the area under
the density curve.
• 𝑃 𝑥1 < 𝑋 < 𝑥2 = 𝐵𝑎𝑠𝑒 × 𝐻𝑒𝑖𝑔ℎ𝑡
f(x)
1
= (𝑥2 − 𝑥1 ) × ( )
𝑏−𝑎
1
𝑏−𝑎
• Or from the distribution function 𝐹 𝑥 :
𝑃 𝑥1 < 𝑋 < 𝑥2 = 𝐹 𝑥2 − 𝐹(𝑥1 ) x
a 𝑥1 𝑥2 b
F(x)
1
F(𝑥2 )
F(𝑥1 )
x
a 𝑥1 𝑥2 b
Example 8.1(a) Uniformly Distributed Gasoline
Sales… 8.37
• The amount of gasoline sold daily at a service station is uniformly
distributed with a minimum of 2,000 gallons and a maximum of
5,000 gallons.
f(x)
2,000 5,000 x
• Find the probability that daily sales will fall between 2,500
and 3,000 gallons.
Algebraically: what is P(2,500 ≤ X ≤ 3,000) ?
Example 8.1(a)…
8.38
P(2,500 ≤ X ≤ 3,000) = (3,000 – 2,500) x = .1667
f(x)
2,000 5,000 x
“There is about a 17% chance that between 2,500 and

3,000 gallons of gas will be sold on a given day.”
Example 8.1(b)…
8.39
distributed with a minimum of 2,000 gallons and a maximum of 5,000
gallons.
f(x)
2,000 5,000 x
• What is the probability that the service station will sell at least
4,000 gallons?
Algebraically: what is P(X ≥ 4,000) ?
Example 8.1(b)…
8.40
P(X ≥ 4,000) = (5,000 – 4,000) x = .3333
f(x)
2,000 5,000 x
“There is a one-in-three chance the gas station will

sell more than 4,000 gallons on any given day.”
Example 8.1(c)…
8.41
distributed with a minimum of 2,000 gallons and a maximum of 5,000
gallons.
f(x)
2,000 5,000 x
• What is the probability that the station will sell exactly 2,500
gallons?
Algebraically: what is P(X = 2,500) ?
Example 8.1(c)…
8.42
P(X = 2,500) = (2,500 – 2,500) x =0
f(x)
2,000 5,000 x
“The probability that the gas station will sell exactly

2,500 gallons is zero.”
Uniform Distribution (균일분포, 일양분포)…
8.43
• Example: Let 𝑋~𝑈(0,1).
a. Graph the pdf.
• Calculate the probability that

b. When 𝐵 = 𝑎, 𝑏 , 𝑤ℎ𝑒𝑟𝑒 𝑎 ≥ 0, 𝑏 ≤ 1, 𝑃 𝑋 ∈ 𝐵 =?
1
c. 𝑃(𝑋 ≤ )=?
2
1 1
d. 𝑃 𝑋 ≤ 𝑋≤ =?
2 4
1 3
e. 𝑃 𝑋 ≤ 𝑋≤ =?
2 4
FYI: Uniform Distribution (균일분포, 일양분포)…
8.44
• Notation: 𝑋~𝑈 𝑎, 𝑏 𝑜𝑟 𝑋~𝑈[𝑎, 𝑏]
• A random variable 𝑋 follows a uniform distribution on the interval (𝑎, 𝑏).
• Definition 8.6 (Parameters of the density function) The constants

that determine the specific form of a density function are called
parameters of the density function.
• The quantities 𝑎 𝑎𝑛𝑑 𝑏 are parameters of the uniform density
function
• Both the range and the probability that 𝑋 will fall in any given
interval depends on the values of 𝑎 𝑎𝑛𝑑 𝑏.
FYI: Uniform Distribution (균일분포, 일양분포) …
8.45
• Theorem 8.6 If 𝑋~𝑈 𝑎, 𝑏 , then
𝑎+𝑏
𝐸 𝑋 = , and
2
2 (𝑏−𝑎)2
𝜎 =𝑉 𝑋 =
12
Proof:
𝑏 𝑏
1 1 𝑥 2 𝑏 𝑏 2 − 𝑎2 𝑎+𝑏
𝐸 𝑋 = න 𝑥𝑓 𝑥 𝑑𝑥 = න 𝑥( ) 𝑑𝑥 = [ ]𝑎 = = ∎
𝑎 𝑎 𝑏−𝑎 𝑏−𝑎 2 2(𝑏 − 𝑎) 2
2 𝑏 𝑏 1 1 𝑥3 𝑏 𝑏3 −𝑎3 𝑎2 +𝑎𝑏+𝑏2
𝐸 𝑋 = ‫ 𝑥 𝑎׬‬2 𝑓 𝑥 𝑑𝑥 = ‫ 𝑥 𝑎׬‬2 ( ) 𝑑𝑥
𝑏−𝑎
= [ ] =
𝑏−𝑎 3 𝑎 3(𝑏−𝑎)
=
3
𝑎2 +𝑎𝑏+𝑏2 𝑎+𝑏 2 (𝑏−𝑎)2

𝑉 𝑋 =𝐸 𝑋2 − 𝐸 𝑋 2 = − = ∎
3 2 12
8-1b Using a Continuous Distribution to
Approximate a Discrete Distribution 8.46
• Discrete or Continuous random variables?
• In practice, we frequently use a continuous distribution to
approximate a discrete one when the number of values the variable
can assume is countable but large.
• Example: the number of possible values of weekly income
• We can easily identify (and thus count) all possible values, so
weekly income is discrete random variable.
• Because it can assume such a large number of values, we prefer to
employ a continuous probability distribution (the normal
distribution) to determine the probability associated with such
variables.
8-2 The Normal Distribution…
8.47
• Abraham de Moivre (26 May 1667-27 November 1754)
• a French mathematician,
published The Doctrine of Chances in 1733,
derived the normal distribution.
• The first who postulate the central limit theorem,

a cornerstone of probability theory.
© 2018 Cengage Learning. All Rights Reserved. May 알수not없는be 작성자

copied,님의
scanned,
이 사진에는or duplicated,
CC BY-SAin라이선스가
whole or in part, except for use as permitted in
적용됩니다.
8.48
• Why is it important?
• Can model numerous random variables that arise in practice:
• heights of people, test scores, rainfall amounts, and scientific
measurements,…
• Provides useful approximation to Binomial and Poisson.
• Basis for statistical inference (Central Limit Theorem: CLT)

• Sampling distribution of mean of every distribution converges to
normal distribution.
• If 𝑋1 , 𝑋2 , … , 𝑋𝑛 are simple random sample with (𝜇, 𝜎 2 ), then as 𝑛
1
increases 𝑋ത → 𝑁(𝜇, 𝜎 2 ) (see chap.9)
𝑛
8.49
• Definition 8.7 (Normal Probability Distribution): A random
variable 𝑋 is said to have a normal probability distribution if and only
if, for 𝜎 > 0 and −∞ < 𝜇 < ∞, the density function of 𝑋 is
1 𝑥−𝜇 2
1 −
𝑓 𝑥 = 𝑒 2 𝜎 , −∞ < 𝑥 < ∞
𝜎 2𝜋
• Notation: 𝑋~𝑁(𝜇, 𝜎 2 )
• Properties:
• Bell-shaped f(x)
• Symmetric around the mean 𝜇
• mean=median=mode
• Empirical rule for 𝜎 (see chap. 4)
x
8.50
• Theorem 8.7 The normal distribution is fully defined by two
parameters: its standard deviation 𝜎 and mean 𝜇.
1 𝑥−𝜇 2
1 −2 𝜎
𝑓 𝑥 = 𝑒 , −∞ < 𝑥 < ∞
𝜎 2𝜋
𝜋 = 3.14159, 𝑒 = 2.71828
• Two parameters: 𝐸 𝑋 = 𝜇 𝑎𝑛𝑑 𝑉 𝑋 = 𝜎 2
• Unlike the range of the uniform distribution (𝑎 ≤ 𝑋 ≤ 𝑏), normal
distribution range from minus infinity to plus infinity.
∞
⇒ ‫׬‬−∞ 𝑓 𝑥 𝑑𝑥 = 1
8-2 Normal Distribution…
8.51
• The normal distribution is described by two parameters: 𝜇, 𝜎
• Increasing the mean shifts the curve to the right…
• Increasing the standard deviation “flattens” the curve…
8-2a Calculating Normal probabilities…
8.52
• The probability that random variable 𝑋 falls in an interval [𝑎, 𝑏] is
given by
1 𝑥−𝜇 2
𝑏 𝑏 1 −
𝑃 𝑎≤𝑋≤𝑏 = ‫𝑥𝑑)𝑥(𝑓 𝑎׬‬ = ‫ 𝜎 𝑎׬‬2𝜋 𝑒 2 𝜎 𝑑𝑥
• Unfortunately, a closed form expression for this integral does not

exist.
• To avoid tedious calculation for every possible normal distribution
with mean 𝜇 and standard deviation 𝜎, we employ Standard Normal
Distribution whose mean is zero and standard deviation is 1.
Standard Normal Distribution…
8.53
• Definition 8.8 (Standard Normal Distribution) A random variable 𝑍
is said to have a standard normal probability distribution if and only if
the density function of 𝑍 is
1
1 − 𝑧2
𝑓 𝑧 = 𝑒 2 , −∞ < 𝑧 < ∞
2𝜋
• A normal distribution whose mean is zero and standard deviation is

one is called the standard normal distribution.
• Notation: 𝑍~𝑁(0,1)
f(z)
• You can think of 𝑧 as a measure of the number of standard deviations 𝑋

𝑥−𝜇
is from 𝜇: 𝑍 =
𝜎
Standard Normal Distribution…
8.54
• Theorem 8.8 If 𝑋 is normally distributed with parameters 𝜇 𝑎𝑛𝑑 𝜎 2 ,
then 𝑌 = 𝑎 + 𝑏𝑋 is normally distributed with mean 𝑎 + 𝑏𝜇 and
𝑣𝑎𝑟𝑖𝑎𝑛𝑐𝑒 𝑏 2 𝜎 2 .
⇔ If 𝑋~𝑁(𝜇, 𝜎 2 ), then 𝑌 = 𝑎 + 𝑏𝑋~𝑁(𝑎 + 𝑏𝜇, 𝑏 2 𝜎 2 )
𝑋−𝜇 1 𝜇
• Example: 𝑋~𝑁(𝜇, 𝜎 2 ), then 𝑍 = = ( 𝑥 − )~𝑁(0,1)
𝜎 𝜎 𝜎
1
𝐸 𝑍 = [𝐸 𝑋 − 𝜇] = 0∎
𝜎
2 2 𝑋−𝜇 2 1 2
𝑉 𝑍 =𝐸 𝑍 − 𝐸 𝑍 =𝐸 = 𝐸 𝑋−𝜇 = 1∎
𝜎 𝜎2
• Any normal distribution can be converted to a standard normal

𝑋−𝜇
distribution with simple algebra. Just use 𝑍 = .
𝜎
Calculating Normal Probabilities…
𝑋−𝜇 8.55
• We can use the function 𝑍 = to convert any normal random
𝜎
variable to a standard normal random variable…
This shifts the mean

of X to zero…
Calculating Normal Probabilities…
𝑋−𝜇
8.56
We can use the function 𝑍 = to convert any normal random
𝜎
variable to a standard normal random variable…
This changes the

shape of the curve…
Example 8.2 Normally Distributed Gasoline
Sales 8.57
• Suppose that at another gas station the daily demand for regular
gasoline is normally distributed with a mean of 1,000 gallons and a
standard deviation of 100 gallons.
• The station manager has just opened the station for business and
notes that there is exactly 1,100 gallons of regular gasoline in storage.
• The next delivery is scheduled later today at the close of business. The
manager would like to know the probability that he will have enough
regular gasoline to satisfy today’s demands.
Sales 8.58
• Let 𝑋 denote the daily gasoline demand. Then 𝑋~𝑁(1,000, 1002 ).
• To find the probability 𝑃 𝑋 < 1,100 , the first step is to standardize 𝑋.
𝑋 − 𝜇 1,100 − 1,000
𝑃 𝑋 < 1,100 = 𝑃 < =𝑃 𝑍<1
𝜎 100
Sales 8.59
• The values of 𝑍 specify the location of the corresponding value of 𝑋.
• A value of 𝑍 = 1 corresponds to a value of 𝑋 that is 1 standard
deviation above the mean.
• Notice as well that the mean of 𝑍, which is 0, corresponds to the
mean of 𝑋.
• If we know the mean and standard deviation of a normally

distributed random variable, we can always transform the probability
statement about 𝑋 into a probability statement about 𝑍.
• Consequently, we need only one table, Table 3 in Appendix B, the

standard normal probability table.
Table 3…
8.60
• This table is similar to the ones we used for the binomial and Poisson
distributions.
• That is, this table lists cumulative probabilities 𝑃(𝑍 < 𝑧) for values of 𝑧
ranging from −3.09 to +3.09.
• Suppose we want to determine the following probability.

𝑃(𝑍 < −1.52)
• We first find −1.5 in the left margin. We then move along this row
until we find the probability under the .02 heading. Thus,
𝑃 𝑍 < −1.52 = .0643
Table 3…
8.61
Table 3…
8.62
𝑃 𝑍 < −1.52 = .0643
Table 3…
8.63
• We can also determine the probability that the standard normal
random variable is greater than some value of 𝑧.
• For example, we find the probability that 𝑍 is greater than 1.80 by

determining the probability that 𝑍 is less than 1.80 and subtracting
that value from 1.
• Applying the complement rule we get
𝑃 𝑍 > 1.80 = 1 − 𝑃 𝑍 < 1.80 = 1 − .9641 = .0359
Table 3…
8.64
𝑃 𝑍 > 1.80 = 1 − 𝑃 𝑍 < 1.80 = 1 − .9641 = .0359
Table 3…
8.65
• We can also easily determine the probability that a standard normal
random variable lies between 2 values of 𝑧.
• For example, we find the probability
𝑃(−1.30 < 𝑍 < 2.10)
• By finding the 2 cumulative probabilities and calculating their
difference:
𝐹 2.10 = 𝑃 𝑍 < 2.10 = .9821,
𝐹 −1.30 = 𝑃 𝑍 < −1.30 = .0968
Hence,
𝑃 −1.30 < 𝑍 < 2.10 = 𝐹 2.10 − 𝐹 −1.30
= 𝑃 𝑍 < 2.10 − 𝑃 𝑍 < −1.30
= .9821 − .0968 = .8853∎
Table 3…
8.66
𝑃 −1.30 < 𝑍 < 2.10 = .8853
Table 3…
8.67
• Notice that the largest value of 𝑧 in the table is 3.09, and that
𝑃 𝑍 < 3.09 = .9990.
• This means that
𝑃 𝑍 > 3.09 = 1 − .9990 = .0010
• However, because the table lists no values beyond 3.09, we

approximate any area beyond 3.10 as 0. That is,
𝑃 𝑍 > 3.09 = 𝑃 𝑍 < −3.09 ≈ 0
Table 3…
8.68
• Recall that in Tables 1 and 2 we were able to use the table to find
the probability that 𝑋 is equal to some value of 𝑥, but we won’t
do the same with the normal table.
• Remember that the normal random variable is continuous and the

probability that a continuous random variable is equal to any
single value is 0.
Example 8.2…
8.69
• Finally returning to Example 8.2, the probability that we seek is
𝑃 𝑋 < 1,100 = 𝑃 𝑍 < 1.00 = .8413
Examples
8.70
• Let your height 𝑋 have a normal distribution with mean 170cm, and
standard deviation 5cm. That is, 𝑋~𝑁(170, 52 ). Find the following
probabilities:
a) 𝑃 165 < 𝑋 < 175 = 𝑃 −1 < 𝑍 < 1 =
b) 𝑃 160 < 𝑋 < 180 = 𝑃 −2 < 𝑍 < 2 =
c) 𝑃 155 < 𝑋 < 185 = 𝑃 −3 < 𝑍 < 3 =
• Check ‘Empirical Rule’:

𝑃 𝜇𝑥 − 𝜎𝑥 ≤ 𝑋 ≤ 𝜇𝑥 + 𝜎𝑥 = .6826
𝑃 𝜇𝑥 − 2𝜎𝑥 ≤ 𝑋 ≤ 𝜇𝑥 + 2𝜎𝑥 = .9544
𝑃 𝜇𝑥 − 3𝜎𝑥 ≤ 𝑋 ≤ 𝜇𝑥 + 3𝜎𝑥 = .9974
APPLICATIONS IN FINANCE: Measuring Risk
8.71
• In Section 7.4 we developed an important application in finance
where the emphasis was placed on reducing the variance of the
returns on a portfolio.
• However, we have not demonstrated why risk is measured by the
variance and standard deviation. The following example corrects
this deficiency.
Example 8.3 Probability of a Negative
Return on Investment 8.72
• Consider an investment whose return is normally distributed with a

mean of 10% and a standard deviation of 5%.
⇔ 𝑋~𝑁(10, 52 )
a. Determine the probability of losing money: 𝑃(𝑋 < 0).
𝑋−𝜇 0−10
𝑃 𝑋<0 =𝑃 < = 𝑃 𝑍 < −2 = .0228 (Table 3)
𝜎 5
b. Find the probability of losing money when the standard deviation is

equal to 10%: 𝑃(𝑋 < 0), when 𝑋~𝑁(10, 102 )
Example 8.3 Probability of a Negative Return
on Investment 8.73
b. If we increase the standard deviation to 10%, the probability of
suffering a loss becomes
𝑋−𝜇 0−10
𝑃 𝑋<0 =𝑃 < = 𝑃 𝑍 < −1 = .1587 (Table 3)
𝜎 10
⇒ Increasing the standard deviation increases the probability of losing

money and also will increase the probability that the return will exceed
some relatively large amount.
• Because investors tend to be risk averse, we emphasize the increased

probability of negative returns when discussing the effect of
increasing the standard deviation.
8-2b Finding Values of 𝑍…
8.74
• Often we’re asked to find some value of 𝑍 for a given probability,
i.e. given an area (𝐴) under the curve, what is the corresponding
value of 𝑧 (𝑧𝐴 ) on the horizontal axis that gives us this area? That
is:
Figure 8.15 𝑃 𝑍 > 𝑧𝐴 = 𝐴*
* 𝑃 𝑍 > 𝑧𝛼 = 𝛼
8.75
• What value of 𝑧 corresponds to an area under the curve of 2.5%? That
is, what is 𝑧.025 ?
• 1 − 𝐴 = 1 − .025 = .9750
• If you do a ‘reverse look-up’ on Table 3 for .9750, you will get the
corresponding 𝑧.025 = 1.96.
• Since 𝑃 𝑍 > 𝑧.025 = .025, we say: 𝑧.025 = 1.96
8.76
• Examples 8.4-5
• Find the value of a standard normal random variable such that the
probability that the random variable is less than it is 5%.
⇒ We want to find 𝑧 value such that 𝑃 𝑍 < 𝑧 = .05.
Since the curve is symmetric about 0, finding 𝑧.05 gives the desired
value as −𝑧.05
From the Table 3, we find 𝑧.05 =1.645. Hence, −𝑧.05 =−1.645.
8-2c 𝑧𝐴 (or 𝑧𝛼 )and Percentiles…
8.77
• Percentiles are measures of relative standing. (see Chap. 4)
• The value of 𝑧𝐴 are 100(1−𝑧𝐴 )𝑡ℎ percentile of a standard normal
random variable.
• 𝑧.05 =1.645 is the 95𝑡ℎ percentile: 95% of all values of 𝑍 are below it
and 5% are above it.
Example 8.6 Determining the Reorder Point…
8.78
• During the spring, the demand for electric fans at a large home-
improvement store is quite strong. The company tracks inventory using
a computer system so that it knows how many fans are in the inventory
at any time.
• The policy is to order a new shipment of 250 fans when the inventory
level falls to the reorder point, which is 150. However, this policy has
resulted in frequent shortages and thus lost sales because both lead
time and demand are highly variable.
• The manager would like to increase reorder point from 150 to some
level to reduce the incidence of shortages and sales loss. The new
policy is that the company reorders when the inventory level reaches to
the 95% of the demand (a 95% service level).
8.79
• From previous periods, the company has determined that demand
during lead time* is normally distributed with a mean of 200 and a
standard deviation of 50.
• Find the reorder point.
Fig. 8.20 Distribution of Demand during Lead Time
.05
200 ROP
* lead time: the time between the beginning of a process or project and the appearance of its results
(상품생산 시작부터 완성까지 걸리는 시간).
8.80
• Solution:
• Let 𝑋 denote the demand for electric fan during lead time* then,
𝑋~𝑁(200, 502 )
• The reorder point is set so that the probability that demand during
lead time exceeds this quantity is 5%. The standardized value of the
reorder point (ROP) is 𝑧.05 = 1.645. To find ROP, we must
unstandardize 𝑧.05 .
𝑅𝑂𝑃−200
• 𝑧.05 =1.645=
50
• 𝑅𝑂𝑃 = 50 1.645 + 200 = 282.25

• Round it up to 283.
Digression: Theorem on Normal Distribution…
8.81
• Theorem 8.9 Let 𝑋1 , 𝑋2 , ⋯ , 𝑋𝑛 be independently normally distributed
random variables with 𝐸 𝑋𝑖 = 𝜇𝑖 𝑎𝑛𝑑 𝑉 𝑋𝑖 = 𝜎𝑖2 , 𝑓𝑜𝑟 𝑖 = 1, 2, … , 𝑛 and
let 𝑎1 , 𝑎2 , … , 𝑎𝑛 be constants. If
𝑈 = σ𝑛𝑖=1 𝑎𝑖 𝑋𝑖 = 𝑎1 𝑋1 + 𝑎2 𝑋2 + ⋯ + 𝑎𝑛 𝑋𝑛 ,
then 𝑈 is a normally distributed random variable with
𝐸(𝑈) = σ𝑛𝑖=1 𝑎𝑖 𝜇𝑖 = 𝑎1 𝜇1 + 𝑎2 𝜇2 + ⋯ + 𝑎𝑛 𝜇𝑛
and
𝑉 𝑈 = σ𝑛𝑖=1 𝑎𝑖2 𝜎𝑖2 = 𝑎12 𝜎12 + 𝑎22 𝜎22 + ⋯ + 𝑎𝑛2 𝜎𝑛2
FYI: Theorem…Examples
8.82
• Example 1: 𝑋𝐴 ~𝑁(𝜇𝐴 , 𝜎𝐴2 ) and 𝑋𝐵 ~𝑁(𝜇𝐵 , 𝜎𝐵2 ) and mutually independent,
then 𝑌 = (𝑋𝐴 − 𝑋𝐵 ) has the following normal distribution:
𝑌~𝑁(𝜇𝐴 − 𝜇𝐵 , 𝜎𝐴2 + 𝜎𝐵2 )
• Example 2: Let 𝑋1 , 𝑋2 , ⋯ , 𝑋𝑛 be a random sample of size 𝑛 from normal

1
distribution, that is, 𝑁(𝜇, 𝜎 2 ). Then 𝑋ത = σ𝑛𝑖=1 𝑋𝑖 has the following
𝑛
normal distribution:
1
ത
𝑋~𝑁 𝜇𝑥ҧ = 𝜇, 𝜎𝑥2ҧ = 𝜎 2
𝑛
ത
𝑋−𝜇 ത
𝑛(𝑋−𝜇)
𝑍= 1 = ~𝑁(0,1)
𝜎 𝜎
𝑛
FYI: Theorem…Examples
8.83
• Theorem 8.10 If 𝑋~𝑁(𝜇𝑋 , 𝜎𝑥2 ) and 𝑌~𝑁(𝜇𝑌 , 𝜎𝑦2 ) and 𝜎𝑥𝑦 = 0, then
𝑋 𝑎𝑛𝑑 𝑌 are independent.
• In normal distribution 𝜎𝑥𝑦 = 0 (𝑜𝑟 𝜌𝑥𝑦 = 0) ⇒ 𝑋 𝑎𝑛𝑑 𝑌 are

independent.
• Generally, 𝜎𝑥𝑦 = 0 𝑜𝑟 𝜌𝑥𝑦 = 0 does not imply that 𝑋 𝑎𝑛𝑑 𝑌 are
independent!!
8.3 (Optional) Exponential Distribution…
8.84
• Definition 8.9 (Exponential Distribution) A random variable 𝑋 is
exponentially distributed if its probability density function is given by
𝑓 𝑥 = 𝜆𝑒 −𝜆𝑥 , 𝑥 ≥ 0
where 𝑒 = 2.71828 … and 𝜆 is the parameter of the distribution.
• Note that 𝑥 ≥ 0.
• Time (for example) is a non-negative quantity; the exponential
distribution is often used for time related phenomena such as the
length of time between phone calls or the life time of electronic
components, etc.
• For the exponential random variable 𝑋,
1
𝐸 𝑋 =σ= ∎
𝜆
8-3 Exponential Distribution…
8.85
• The exponential distribution depends upon the value of 𝜆.
𝑓 𝑥 = 𝜆𝑒 −𝜆𝑥 , 𝑥 ≥ 0
• Example: exponential distributions for 𝜆 = 0.5, 1, 2)
𝑓 0 = 𝜆: Smaller values of 𝜆 “flatten” the curve
FYI: Exponential Distribution…
8.86
• 𝑓 𝑥 = 𝜆𝑒 −𝜆𝑥 , 𝑥 ≥ 0
𝑎 𝑎 𝑎
• 𝑃 𝑋<𝑎 =𝐹 𝑎 = ‫׬‬0 𝑓 𝑥 𝑑𝑥 = ‫׬‬0 𝜆𝑒 −𝜆𝑥 𝑑𝑥 = −𝑒 −𝜆𝑥
0
= −𝑒 −𝜆𝑎 − − 1 = 1 − 𝑒 −𝜆𝑎 ∎
𝐹 ∞ = lim 𝐹 𝑎 = lim {1 − 𝑒 −𝜆𝑎 ] = 1∎
𝑎→∞ 𝑎→∞
• 𝑃 𝑋 > 𝑎 = 1 − 𝐹 𝑎 = 1 − 1 − 𝑒 −𝜆𝑎 = 𝑒 −𝜆𝑎 ∎

• 𝑃 𝑥1 < 𝑋 < 𝑥2 = 𝐹 𝑥2 − 𝐹 𝑥1 = [1 − 𝑒 −𝜆𝑥2 ] − [1 − 𝑒 −𝜆𝑥1 ]
= 𝑒 −𝜆𝑥1 − 𝑒 −𝜆𝑥2 ∎
Example 8.7 Lifetimes of Alkaline
Batteries… 8.87
• The lifetime of an alkaline battery (measured in hours) is exponentially
distributed with 𝜆 = .05.
a. What is the mean and standard deviation of the battery’s lifetime?
1 1
𝜇=𝜎= = = 20 ℎ𝑜𝑢𝑟𝑠∎
𝜆 .05
b. Find the probability that a battery will last between 10 and 15 hours.
𝑃 10 < 𝑋 < 15 = 𝑒 −.05(10) − 𝑒 −.05 15 = .6065 − .4724 = .1341∎
c. What is the probability that a

battery will last for more
than 20 hours?
𝑃 𝑋 > 20 = 𝑒 −.05(20) = .3679∎
Example 8.8 Supermarket Checkout
Counter… 8.88
• A checkout counter at a supermarket completes the process according
to an exponential distribution with a service rate of 6 per hour. A
customer arrives at the checkout counter. Find the probability of the
following events.
• To solve this problem is to determine a service rate per minute.
6
That is, 𝜆 = = .1/min. or
60
a. 𝑃 𝑋 < 5 = 1 − 𝑒 −𝜆𝑥 = 1 − 𝑒 − .1 5 = 1 − 𝑒 −.5 = 1 − .6065 = .3935

b. 𝑃 𝑋 > 10 = 𝑒 −𝜆𝑥 = 𝑒 − .1 10 = 𝑒 −1 = .3679
c. 𝑃 5 < 𝑋 < 8 = 𝑒− .1 5 − 𝑒− .1 8 = .6065-.4493=.1572
5 10
[[or 𝜆 = 6/hr a. 𝑋 < ; b. 𝑋 > ; c. 5 < 𝑋 < 8
60 60
5
5
𝑃 𝑋< =1− 𝑒 −𝜆𝑥 =1− 𝑒 − 6 60 = 1 − 𝑒 −.5 ]]
60
8-4 Other Continuous Distributions…
8.89
• Three other important continuous distributions which will be used
extensively in later sections are introduced here:
• Student 𝑡 Distribution,
• Chi-Squared Distribution, and
• 𝐹 Distribution.
8-4a Student 𝑡 Distribution…
8.90
• First derived by William S. Gosset in 1908.
• (13 June 1876 – 16 October 1937)
• an English statistician, chemist and brewer
• Gosset published his findings under the pseudonym “Student” and

used the letter 𝑡 to represent the random variable, hence the Student
𝑡 distribution—also called the Student’s 𝑡 distribution.)
알 수 없는 작성자 님의 이 사진에는 CC BY-NC-ND 라이선스가 적용됩니다.
Student 𝑡 Distribution…
8.91
• Definition 8.10 (Student t distribution) Here the letter 𝑡 is used
to represent the random variable, hence the name. The density
function for the Student 𝑡 distribution is as follows…
𝜈+1
Γ( ) 𝑡 2 −(𝜈+1)
𝑓(𝑡)= 𝜈
2
[1 + ] 2 , ∞ < 𝑡 < ∞(*)
Γ(2) 𝜈𝜋 𝜈
where 𝜈 (nu) is called the degrees of freedom and

Γ(𝑘)(Gamma function) is
Γ 𝑘 = 𝑘−1 𝑘−2 ⋯ 2 1 = 𝑘−1 !
𝑛+1
Γ( ) 𝑡 2 −(𝑛+1)
* non-Greek expression : 𝑓(𝑡)= 𝑛
2
[1 + ] 2
Γ( ) 𝑛𝜋 𝑛
2
** 𝜋: the ratio of the circumference of a circle to its diameter (= 3.14159265…)
8.92
• Much like the standard normal distribution, the Student 𝑡 distribution is
“mound” shaped and symmetrical about its mean of zero:
• The mean and variance of a Student 𝑡 random variable are

𝐸 𝑡 =0
𝜈 𝑛
𝑉(𝑡)= (> 1) for 𝜈 > 2 [or 𝑉 𝑡 = 𝑓𝑜𝑟 𝑛 > 2]
𝜈−2 𝑛−2
8.93
• In much the same way that 𝜇 and 𝜎 define the normal distribution,
𝜈 (or 𝑛), the degrees of freedom, defines the Student 𝑡 Distribution:
𝜈
• 𝑉 𝑡 =
𝜈−2
• 𝜈 = 10, 𝑉 𝑡 = 1.25
• 𝜈 = 50, 𝑉 𝑡 = 1.042
• 𝜈 = 200, 𝑉 𝑡 = 1.010
• As the number of degrees of freedom increases, the 𝑡 distribution

approaches the standard normal distribution.
Student t Probabilities
8.94
• For each value of 𝜈 (the number of degrees of freedom), there is a
different Student t distribution. → We need a different distribution table
to calculate probabilities of Student t random variable manually.
• Not practical.
• We can use Microsoft Excel. (p.277)
• To compute Student t probabilities, type
=TDIST([x], [𝜈], [Tails])
where x must be positive, 𝜈 is the number of degrees of freedom, and “Tails” 1
or 2.
• Typing 1 for “Tails” produces the area to the right of x. Typing 2 for “Tails”
produces the area to the right of x plus the area to the left of –x.
Determining Student 𝑡 Values…
8.95
• The student 𝑡 distribution is used extensively in statistical inference.
• Table 4 in Appendix B lists values of 𝑡𝐴,𝜈 .
• Values of a Student 𝑡 random variable with 𝜈 degrees of freedom such that
𝑃 𝑡 > 𝑡𝐴,𝜈 = 𝐴. (Fig. 8.26)
Fig. 8.26 Student t distribution with 𝑡𝐴
Determining Student 𝑡 Values…
8.96
• Table 8.2 (or Table 4 in Appendix B) lists values of 𝑡𝐴,𝜈 .
• The values for 𝐴 are pre-determined “critical” values, typically in the
10%, 5%, 2.5%, 1% and 1/2% range.
• Observe that 𝑡𝐴,𝜈 is provided for degrees of freedom ranging from 1
to 200 and ∞.
• To read this table, simply identify the degrees of freedom and find
that value or the closest number to it if it is not listed.
• Then locate the column representing the 𝑡𝐴 value you wish.
Using the 𝑡 table for values…
8.97
• For example, if we want the value of 𝑡 with 10 degrees of
freedom such that the area under the Student 𝑡 curve is .05:
Area under the curve value (𝑡𝐴 ) : COLUMN
t.05,10
t.05,10=1.812
Degrees of Freedom : ROW
8.98
• If the number of degrees of freedom is not shown, find its closest
value.
• For example, suppose we wanted to find 𝑡.025,32 . Because 32
degrees of freedom is not listed, we find the closest number of
degrees of freedom, which is 30, and use 𝑡.025,30 = 2.042 as an
approximation.
• Because the Student 𝑡 distribution is symmetric about 0, the value

of 𝑡 such that the area to its left is 𝐴 is −𝑡𝐴,𝜈 .
• For example, the value of 𝑡 with 10 degrees of freedom such
that the area to its left is .05 is
−𝑡.05,10 = −1.812
8.99
• Notice the last row in the Student 𝑡 table. The number of degrees of
freedom is infinite, and the 𝑡 values are identical (except for the
number of decimal places) to the values of 𝑧.
𝑡.10,∞ = 1.282 𝑧.10 = 1.28

𝑡.05,∞ = 1.645 𝑧.05 = 1.645
𝑡.025, ∞ = 1.960 𝑧.025 = 1.96
𝑡.01,∞ = 2.326 𝑧.01 = 2.23
𝑡.005,∞ = 2.576 𝑧.005 = 2.575
8-4b Chi-Squared Distribution…
8.100
• Definition 8.11 (Chi-Squared Distribution) The chi-squared density
function is given by:
𝜈 −𝜒2
1 1
𝑓 𝜒2 = 𝜈 (𝜒 2 )2−1 𝑒 2 , 𝜒 2 > 0.
Γ( ) 2𝜈/2
2
• The parameter 𝜈 is the number of degrees of freedom, which like the

degrees of freedom of the Student 𝑡 distribution affects the shape.
• Chi-squared distribution is positively skewed ranging between 0 and
∞.
8-4b Chi-Squared Distribution… 8.10
1
• The effects of increasing the degrees of freedom is seen in Figure
8.28.
• The mean and variance of a chi-squared random variable are:
𝐸 𝜒 2 = 𝜈, and
𝑉 𝜒 2 = 2ν
• 𝜒 2 > 0, its distribution is not symmetrical.

• When 𝜈 > 100, the chi-squared distribution can be approximated by a
normal distribution with
𝜇 = 𝜈 𝑎𝑛𝑑 𝜎 = 2𝜈
• Determining Chi-squared values:
• The value of 𝜒 2 with 𝜈 degrees of freedom such that the area to its
right under the chi-squared curve is equal to 𝐴 is denoted by 𝜒𝐴,𝜈
2
.
2
• 𝑃(𝜒 2 > 𝜒𝐴,𝜈 )=𝐴
• To represent left-tail critical values, 𝜒1−𝐴,𝜈
2
denotes
the point such that the area to its left is 𝐴.
2
• 𝑃 𝜒 2 < 𝜒1−𝐴,𝜈 =𝐴
2
𝜒1−𝐴,𝜈 𝜒2
For Example… 8.105
• To find the point in a chi-squared distribution with 8 degrees of freedom,
such that the area to the right is .05:
• Locate 8 degrees of freedom in the left column and 𝜒.05
2
across the
top. The intersection of the row and column contains the number we
seek.
2
• 𝜒.05,8 =15.5
For Example… 8.106
• To find the point in a chi-squared distribution with 8 degrees of
freedom, such that the area to the left is .05,
• Look up the intersection of the 8 degrees of freedom with the
column 𝜒.950
2
, yielding a value of 2.73
2
𝜒.950
8-4c 𝐹 Distribution… 8.107
• Definition 8.12 (F Distribution) The density function of 𝐹

distribution is given by:
𝜈 +𝜈 𝜈1 𝜈1 −2
Γ( 1 2 2) 𝜈1 2 𝐹 2
𝑔 𝐹 = 𝜈 𝜈 𝜈 +𝜈 , 𝐹>0
Γ( 21 )Γ( 22 ) 𝜈2 𝜈1 𝐹 1 2
(1+ 𝜈 ) 2
2
• 0<𝐹<∞
• 𝜈1 and 𝜈2 are the parameters of the distribution and are called the
degrees of freedom
• 𝜈1 : the numerator degrees of freedom
• 𝜈2 : the denominator degrees of freedom
8-4c 𝐹 Distribution… 8.108
• The mean and variance of an 𝐹 random variable are given by:

𝜈2
𝐸 𝐹 = , 𝜈2 > 2
𝜈2 −2
And
2𝜈22 (𝜈1 +𝜈2 −2)
𝑉 𝐹 = , 𝜈2 > 4
𝜈1 𝜈2 −2 2 (𝜈2 −4)
• Notice that the mean depends only on the denominator degrees of

freedom and for large 𝜈2 the mean of the distribution is
approximately 1.
• The F distribution is positively skewed .
Determining Values of 𝐹… 8.109
• We define 𝐹𝐴,𝜈1𝜈2 as the value of 𝐹 with 𝜈1 𝑎𝑛𝑑 𝜈2 degrees of

freedom such that the area to its right under the curve is 𝐴; that is,
𝑃 𝐹 > 𝐹𝐴,𝜈1 𝜈2 = 𝐴
• Because the 𝐹 random variable can take on only positive values, we

define 𝐹1−𝐴,𝜈1𝜈2 as the value of 𝐹 with 𝜈1 𝑎𝑛𝑑 𝜈2 degrees of freedom
such that the area to its left under the curve is 𝐴;
𝐹1−𝐴
• For areas under the curve on the left hand side of the curve, we
can leverage the following relationship:
1
𝐹1−𝐴,𝜈1,𝜈2 =
𝐹𝐴,𝜈2 ,𝜈1
• Pay close attention to the order of the terms
• Theorem about 𝐹 distribution:

1
If 𝑋~𝐹 𝑚, 𝑛 , 𝑡ℎ𝑒𝑛 ~𝐹 𝑛, 𝑚 .
𝑋
• (Determining any critical value)

• Values of 𝐹𝐴 such that 𝑃 𝐹 > 𝐹𝐴 = 𝐴 are given in the Table 6,
Appendix B, for values of 𝐴 =.050, .025, .010, and .005.
• In Table 6, the column headings are the numerator degrees of

freedom, whereas the denominator degrees of freedom are given
in the main row headings.
• For example, what is the value of 𝐹 for 5% of the area under the
right hand “tail” of the curve, with a numerator degree of freedom of
3 and a denominator degree of freedom of 7?
Solution: use the 𝐹 table
There are different tables

for different values of A.
Make sure you start with
the correct table!!
F.05,3,7=4.35
F.05,3,7
Denominator Degrees of Freedom : ROW
Numerator Degrees of Freedom : COLUMN
8.113
The End of Chap. 8

SME11e PPT ch08std

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SME11e PPT ch08std

Uploaded by

Copyright:

Available Formats

Slides by:

• Probability Density Functions (확률밀도함수)

• Normal Distribution (정규분포)

• Exponential Distribution (지수분포)

• Other Continuous Distributions

• In Chapter 8, we introduce continuous probability distributions, which

• Consequently, we can determine the probability of only a range of

Fig.8.1 Histogram for Example 3.1

• The probability that an ACBL member is between 25 and 45 years of

* (Cumulative) distribution function: (누적)분포함수

• Solution: The probability function for 𝑋 is given by

• Because the only values of 𝑋 that are assigned to positive

3. 𝐹(𝑥) is a non-decreasing function of 𝑥 (If 𝑥1 𝑎𝑛𝑑 𝑥2 are any

wherever the derivative* exists, called the probability density function

• Theorem 8.2 (Properties of a Density Function 𝑓(𝑥))

* The second fundamental theorem of calculus. See theorem 8.4 in FYI.

Q: Find the probability density function for 𝑥, 1

• Showing three of infinitely many solutions that can be

• Find 𝐹 𝑥 𝑎𝑛𝑑 graph 𝐹 𝑥 .

• As numerical descriptive measures, we need to find means

provided that the integral exists.

• The expected value of function 𝑔 𝑋 = (𝑋 − 𝜇)2 is the

• Notation: 𝑋~𝑈(𝑎, 𝑏) or 𝑋~𝑈[𝑎, 𝑏]

P(2,500 ≤ X ≤ 3,000) = (3,000 – 2,500) x = .1667

“There is about a 17% chance that between 2,500 and

P(X ≥ 4,000) = (5,000 – 4,000) x = .3333

“There is a one-in-three chance the gas station will

“The probability that the gas station will sell exactly

• Calculate the probability that

• Definition 8.6 (Parameters of the density function) The constants

𝑎2 +𝑎𝑏+𝑏2 𝑎+𝑏 2 (𝑏−𝑎)2

• The first who postulate the central limit theorem,

© 2018 Cengage Learning. All Rights Reserved. May 알수not없는be 작성자

• Basis for statistical inference (Central Limit Theorem: CLT)

• Unfortunately, a closed form expression for this integral does not

• A normal distribution whose mean is zero and standard deviation is

• You can think of 𝑧 as a measure of the number of standard deviations 𝑋

• Any normal distribution can be converted to a standard normal

This shifts the mean

This changes the

• If we know the mean and standard deviation of a normally

• Consequently, we need only one table, Table 3 in Appendix B, the

• Suppose we want to determine the following probability.

• For example, we find the probability that 𝑍 is greater than 1.80 by

• Applying the complement rule we get

𝑃 𝑍 > 1.80 = 1 − 𝑃 𝑍 < 1.80 = 1 − .9641 = .0359

• However, because the table lists no values beyond 3.09, we

• Remember that the normal random variable is continuous and the

• Check ‘Empirical Rule’:

• Consider an investment whose return is normally distributed with a

b. Find the probability of losing money when the standard deviation is

⇒ Increasing the standard deviation increases the probability of losing

• Because investors tend to be risk averse, we emphasize the increased

Fig. 8.20 Distribution of Demand during Lead Time

• 𝑅𝑂𝑃 = 50 1.645 + 200 = 282.25

• Example 2: Let 𝑋1 , 𝑋2 , ⋯ , 𝑋𝑛 be a random sample of size 𝑛 from normal

• In normal distribution 𝜎𝑥𝑦 = 0 (𝑜𝑟 𝜌𝑥𝑦 = 0) ⇒ 𝑋 𝑎𝑛𝑑 𝑌 are

𝑓 0 = 𝜆: Smaller values of 𝜆 “flatten” the curve

• 𝑃 𝑋 > 𝑎 = 1 − 𝐹 𝑎 = 1 − 1 − 𝑒 −𝜆𝑎 = 𝑒 −𝜆𝑎 ∎

c. What is the probability that a

a. 𝑃 𝑋 < 5 = 1 − 𝑒 −𝜆𝑥 = 1 − 𝑒 − .1 5 = 1 − 𝑒 −.5 = 1 − .6065 = .3935

• Gosset published his findings under the pseudonym “Student” and