Module 3: Random Variables

Lecture 6: Some standard Continuous Probability distributions

Continuous Random Variables
A continuous random variable is a function that can take any real value within a given range and
its probability density function (pdf) is a continuous over the range. The pdf gives the probability
of a value of the random variable falling within a particular interval.
Probability Density Function (pdf)
A Probability density function (pdf) is conventionally denoted by ) (x f
X
and its properties are
- 0 ) ( > x f
X
for all X , and
-
}

=1 ) ( dx x f
X

Uniform Distribution
Any RV X is a uniformly distributed random variable if its probability density function is given
by

s s

=
elsewhere
x
x f
x
0
1
) , ; (
| o
o |
| o
The cumulative distribution function is given by ( ) | o
o |
o
s s

= x
x
X F . The following
figure (Fig. 1) shows the pdf of a uniform distribution.



Fig. 1. pdf of a uniform probability distribution

Mean, Variance and Coefficient of Skewness of Uniform Distribution
- The mean is given by
2
| o

+
=
- The variance is given by
( )
12
2
2
o |
o

=
- The coefficient of skewness is zero as the distribution is symmetric about the mean
An example
In the context of Civil Engineering, the application of uniform distribution may be understood in
the light of the following examples.
Let us consider a structural member (as shown in the Fig 2a) made up of a material with uniform
characteristics and subject to a particular load. It can fail (break or rupture) at a distance x from
one end. If the probability of failure anywhere along the length is equal, then the probability
distribution of X is uniform with support 0toL . The pdf (shown in Fig 2b) is given by
( )
elsewhere
L x for
L
x f
X
0
0
1
=
s s =

Fig 2a. Location of failure in a structural member
Fig 2b. pdf of uniform distribution of L (location of failure in a structural member)
Similarly, the location of road accident on a highway stretch identified to be accident prone
follows uniform distribution if the occurrence of the accident is equally likely anywhere on the
stretch.
Normal or Gaussian Distribution
The most important and widely used pdf is the Normal Distribution (aka Gaussian Distribution).
It is a continuous probability with unbounded support and symmetrical distribution about mean.
Its two-parameters are mean ( ) and variance ( )
2
o . The pdf of Gaussian distribution with
parameters (mean) and
2
o (variance) is expressed as
( ) < < =
|
.
|

\
|

x e x f
x
2
2
1
2
2
1
, ;
o

o t
o (1)
The pdf results in a bell shaped curve symmetric about the mean (Fig 3a).
The cumulative distribution function is given by ( ) dx e x F
x
x
X
2
2
1
2
1
|
.
|

\
|

}
=
o

o t
. The shape of the
CDF is shown in Fig 3b.


-3 -2 -1 0 1 2 3 4 5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
x
f
X

(
x
)
mu=1, sigma=1.2

Fig 3a. pdf of normal distribution
x
( ) x f
X
-3 -2 -1 0 1 2 3 4 5 6 7 8
0
0.5
1
1.5
x
F
X

(
x
)
mu=1,sigma=1.2

Fig 3b. CDF of normal distribution
The Normal Distribution is a two-parameter distribution. Mean is the shape parameter of the
normal distribution and is generally denoted by . Variance is the scale parameter of the normal
distribution and is generally denoted by
2
o

. The coefficient of skewness is zero as the normal
distribution is symmetric about the mean. The following figures (Fig 3c and Fig 3d) show the
shapes of the normal pdf for different values of the parameters.
-4 -3 -2 -1 0 1 2 3 4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
x
f
X

(
x
)
mu=0, sigma=1
mu=0, sigma=1.5
mu=0, sigma=0.75

Fig 3c. pdf of normal distribution with same mean and different variances
x
( ) x f
X
( ) x f
X
x
-6 -4 -2 0 2 4 6
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
x
f
X

(
x
)
mu=-2, sigma=1.5
mu=0, sigma=1.5
mu=2, sigma=1.5

Fig 3d. pdf of normal distribution with same variance and different means
Application of Normal distribution
In the context of Civil Engineering, a simple application of normal distribution is the crushing
strength of concrete cubes. If a number of concrete cubes, prepared through identical methods
and cured under identical circumstances, are tested for their crushing strength, it is observed that
their crushing strength is normally distributed RV. The crushing strength that is available in at
least % 95 of the samples is called the characteristic strength or the % 95 dependable strength.
Additive Properties of Normally Distributed RVs
If a RV X is normally distributed with parameters and
2
o , then another RV bX a Y + = is
also normally distributed with parameters b a + and
2 2
o b .
If ( ) n i for X
i
,... 2 , 1 = are n independent and normally distributed RVs with parameters
i
and
2
i
o , then the RV
n n
X b X b X b a Y + + + + = ...
2 2 1 1
is normally distributed with parameters

=
+ =
n
i
i i Y
b a
1
and

=
=
n
i
i i Y
b
1
2 2 2
o o .




x
( ) x f
X
Standard Normal Distribution
The normal distribution having mean 0 = and variance 1
2
= o is called the Standard Normal
Distribution. It is given by ( ) < < =

z e z p
z
Z
2
2
2
1
t
. The CDF of standard normal
distribution is given by ( ) dz e z Z P
z z
2
2
2
1


}
= s
t
.Fig 4 shows the standard normal pdf.
-4 -3 -2 -1 0 1 2 3 4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
x
f
X

(
x
)
mu=0, sigma=1

Fig 4. Standard normal pdf

Utility of Standard Normal Distribution
If a RV X is normally distributed with mean and variance
2
o , then the probability of X being
less than or equal to x is given by ( ) dx e x X P
x
x
2
2
1
2
1
|
.
|

\
|

}
= s
o

o t

As the above integration cannot be evaluated analytically, numerically computed values are
tabulated (taking 0 = and 0
2
= o ) in the standard normal table (Table 1). For normal
distributions with 0 = and 1
2
= o , the cumulative probability is determined by converting the
RV X to the reduced variate
o

=
X
Z .

x
( ) x f
X


Table 1. Standard normal probabilities (source: http://www.stat.purdue.edu/~mccabe
/ips4tab/bmtables.pdf)


z .00 .01 .02 .03 .04 .05 .06 .07 .08 .09
-3.4
-3.3
-3.2
-3.1
-3.0
.0003 .0003 .0003 .0003 .0003 .0003 .0003 .0003 .0003 .0002
.0005 .0005 .0005 .0004 .0004 .0004 .0004 .0004 .0004 .0003
.0007 .0007 .0006 .0006 .0006 .0006 .0006 .0005 .0005 .0005
.0010 .0009 .0009 .0009 .0008 .0008 .0008 .0008 .0007 .0007
.0013 .0013 .0013 .0012 .0012 .0011 .0011 .0011 .0010 .0010
-2.9
-2.8
-2.7
-2.6
-2.5
.0019 .0018 .0018 .0017 .0016 .0016 .0015 .0015 .0014 .0014
.0026 .0025 .0024 .0023 .0023 .0022 .0021 .0021 .0020 .0019
.0035 .0034 .0033 .0032 .0031 .0030 .0029 .0028 .0027 .0026
.0047 .0045 .0044 .0043 .0041 .0040 .0039 .0038 .0037 .0036
.0062 .0060 .0059 .0057 .0055 .0054 .0052 .0051 .0049 .0048
-2.4
-2.3
-2.2
-2.1
-2.0
.0082 .0080 .0078 .0075 .0073 .0071 .0069 .0068 .0066 .0064
.0107 .0104 .0102 .0099 .0096 .0094 .0091 .0089 .0087 .0084
.0139 .0136 .0132 .0129 .0125 .0122 .0119 .0116 .0113 .0110
.0179 .0174 .0170 .0166 .0162 .0158 .0154 .0150 .0146 .0143
.0228 .0222 .0217 .0212 .0207 .0202 .0197 .0192 .0188 .0183
-1.9
-1.8
-1.7
-1.6
-1.5
.0287 .0281 .0274 .0268 .0262 .0256 .0250 .0244 .0239 .0233
.0359 .0351 .0344 .0336 .0329 .0322 .0314 .0307 .0301 .0294
.0446 .0436 .0427 .0418 .0409 .0401 .0392 .0384 .0375 .0367
.0548 .0537 .0526 .0516 .0505 .0495 .0485 .0475 .0465 .0455
.0668 .0655 .0643 .0630 .0618 .0606 .0594 .0582 .0571 .0559
-1.4
-1.3
-1.2
-1.1
-1.0
.0808 .0793 .0778 .0764 .0749 .0735 .0721 .0708 .0694 .0681
.0968 .0951 .0934 .0918 .0901 .0885 .0869 .0853 .0838 .0823
.1151 .1131 .1112 .1093 .1075 .1056 .1038 .1020 .1003 .0985
.1357 .1335 .1314 .1292 .1271 .1251 .1230 .1210 .1190 .1170
.1587 .1562 .1539 .1515 .1492 .1469 .1446 .1423 .1401 .1379
-0.9
-0.8
-0.7
-0.6
-0.5
.1841 .1814 .1788 .1762 .1736 .1711 .1685 .1660 .1635 .1611
.2119 .2090 .2061 .2033 .2005 .1977 .1949 .1922 .1894 .1867
.2420 .2389 .2358 .2327 .2296 .2266 .2236 .2206 .2177 .2148
.2743 .2709 .2676 .2643 .2611 .2578 .2546 .2514 .2483 .2451
.3085 .3050 .3015 .2981 .2946 .2912 .2877 .2843 .2810 .2776
-0.4
-0.3
-0.2
-0.1
-0.0
.3446 .3409 .3372 .3336 .3300 .3264 .3228 .3192 .3156 .3121
.3821 .3783 .3745 .3707 .3669 .3632 .3594 .3557 .3520 .3483
.4207 .4168 .4129 .4090 .4052 .4013 .3974 .3936 .3897 .3859
.4602 .4562 .4522 .4483 .4443 .4404 .4364 .4325 .4286 .4247
.5000 .4960 .4920 .4880 .4840 .4801 .4761 .4721 .4681 .4641

It may be noted from the standard normal distribution curve that % 74 . 99 of the area under the
curve falls inside the region bounded by o 3 . This is particularly important in real life
scenarios where a RV may be bounded by 0 = X but can still be considered to be normally
distributed if o 3 > .



Probability Calculation
Let us consider a RV X that is normally distributed with mean 35 . 4 = and standard deviation
59 . 0 = o . Now, the probability that the RV will take a value between 4 and 5 can be
conveniently calculated from the cumulative standard normal probability table.
( ) ( ) ( )
5867 . 0
2776 . 0 8643 . 0
59 . 0 1 . 1
59 . 0
35 . 4 4
59 . 0
35 . 4 5
5 4
=
=
= |
.
|

\
|
|
.
|

\
|
= s s F F F F X P

The required probability is actually equal to the shaded area in the standard normal pdf as shown
below (Fig 5).

Fig 5. Shaded area representing probability in standard normal curve

Log Normal Distribution
Any RV X is a Log Normal random variable if its probability density function is given by
( ) < < =
|
.
|

\
|

x e
x
x f
x
2
ln
2
1
2
2
1
, ;
o

t o
o
The cumulative distribution function is given by ( )
|
.
|

\
|
u =
o

o
x
x F
ln
, ;
2
where u =standard
normal CDF. If the RV X is log normally distributed, then the RV ( ) X Y log = is normally
distributed.
x
( ) x f
X
Fig 6a and Fig 6b show the pdf and CDF of log normal distribution.
0 1 2 3 4 5 6 7 8
0
0.05
0.1
0.15
0.2
0.25
x
f
X

(
x
)
log mu=1,log sigma=1.2

Fig 6a. pdf of log normal distribution
0 1 2 3 4 5 6 7 8 9 10
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x
F
X

(
x
)
log mu=1,log sigma=1.2

Fig 6b. CDF of log normal distribution

Mean, Variance, Coefficient of Variation and Skewness of Log Normal Distribution
Mean is given by ( )
2
2
1
o

+
= = e X E
X

x
x
( ) x f
X
( ) x f
X
Variance is given by
( ) ( ) 1
2
2 2
= =
o
o e X Var
X X

Coefficient of Variation is given by
( ) 1
2
,
=
o
e C
X V

Coefficient of Skewness is given by
3
, ,
3
X V X V X
C C + = . The distribution is positively skewed
and with decrease of the coefficient of variation, the skewness also decreases.
Sample Estimates of parameters of Log Normal distribution
The sample mean for ( ) X Y log = is given by
|
|
.
|

\
|
+
=
2
,
2
1
ln
2
1
X V
C
X
Y
The sample variance for ( ) X Y log = is given by ( )
2
,
2
1 ln
X V Y
C S + = where
X
S
C
X
X V
=
,

Application of Log Normal distribution
In the context of Civil Engineering, distribution of annual river flow data is often expressed as
log normal distribution. This is because the streamflow values are always non zero and the
probability densities for extremely low (or high) streamflows (observed only during droughts or
floods respectively) are quite less. The probability densities are high for moderate values of
streamflows (observed through the greater part of the year) and decreases progressively for
increasing streamflows.
Exponential Distribution
It is the probability distribution that describes the time between events in an experiment where
outcomes occur continuously and independently at a constant average rate. It is generally used as
a decay function in engineering problems.
A random variable X is an exponential RV with parameter , if its probability density function
is given by

<
>
=

0 0
0
) ; (
x
x e
x f
x
x

The cumulative distribution function is given by

<
>
=

0 0
0 1
) ; (
x
x e
x F
x
x

The pdf and CDF of exponential distribution are shown in Fig 7a and Fig 7b respectively.

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x
f
X

(
x
)
lamda=1

Fig. 7a. pdf of exponential distribution

0 1 2 3 4 5 6 7 8
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
x
F
X

(
x
)
lamda=1

Fig. 7b. CDF of exponential distribution


x
( ) x f
X
x
( ) x f
X
Mean, Variance and skewness of Exponential Distribution
The mean is given by | |

1
= X E
The variance is given by | |
2
1

= X Var
The skewness is 2.
Application of Exponential distribution
Applications of exponential distribution are widespread in Civil Engineering. For example, the
random variable daily rainfall depth follows exponential distribution as the probability density
is highest for zero rainfall depth and it becomes progressively lesser for higher values of rainfall
depth.