Continuous random variables

Andrew Thangaraj

IIT Madras

Andrew Thangaraj (IIT Madras) Continuous random variables 1 / 27

 Subsection 1

Introduction

Andrew Thangaraj (IIT Madras) Continuous random variables 2 / 27

Why continuous random variables?
Consider a discrete random variable X taking values in an alphabet X
Suppose that |X | is growing very large and unwieldy for calculations

Andrew Thangaraj (IIT Madras) Continuous random variables 3 / 27

Why continuous random variables?
Consider a discrete random variable X taking values in an alphabet X
Suppose that |X | is growing very large and unwieldy for calculations
Example 1
I Metoerite weight
F 0.01 grams to 60 tons
F Data available for 45000+ meteorites
F Weights spread out over a large range

Andrew Thangaraj (IIT Madras) Continuous random variables 3 / 27

Why continuous random variables?
Consider a discrete random variable X taking values in an alphabet X
Suppose that |X | is growing very large and unwieldy for calculations
Example 1
I Metoerite weight
F 0.01 grams to 60 tons
F Data available for 45000+ meteorites
F Weights spread out over a large range

Example 2
I Binomial(n, p)
F n grows very large
F p is a constant
F can happen in Bernoulli trials

Andrew Thangaraj (IIT Madras) Continuous random variables 3 / 27

Why continuous random variables?
Consider a discrete random variable X taking values in an alphabet X
Suppose that |X | is growing very large and unwieldy for calculations
Example 1
I Metoerite weight
F 0.01 grams to 60 tons
F Data available for 45000+ meteorites
F Weights spread out over a large range

Example 2
I Binomial(n, p)
F n grows very large
F p is a constant
F can happen in Bernoulli trials

Is there a way to simplify the descriptions of these random-like

phenomena?
I Yes! But we need to give up something. . .

Andrew Thangaraj (IIT Madras) Continuous random variables 3 / 27

Meteorite data
Preprocessing: Take logarithms (log2 )
I Range: from [0.01, 60000000] to [−6.6, 25.8]
I Still, we have 45000+ data

Andrew Thangaraj (IIT Madras) Continuous random variables 4 / 27

Meteorite data
Preprocessing: Take logarithms (log2 )
I Range: from [0.01, 60000000] to [−6.6, 25.8]
I Still, we have 45000+ data

Main idea: Move from individual values to intervals of values

I Divide [−6.6, 25.8] into ≈ 100 intervals
F [−6.6, −6.3], [−6.3, −6], . . . , [25.5, 25.8]
I Count the number of values falling inside each interval

Andrew Thangaraj (IIT Madras) Continuous random variables 4 / 27

Meteorite data
Preprocessing: Take logarithms (log2 )
I Range: from [0.01, 60000000] to [−6.6, 25.8]
I Still, we have 45000+ data

Main idea: Move from individual values to intervals of values

I Divide [−6.6, 25.8] into ≈ 100 intervals
F [−6.6, −6.3], [−6.3, −6], . . . , [25.5, 25.8]
I Count the number of values falling inside each interval

Histogram of log of weights

Describe shape of the
histogram instead of individual
values
What do we give up? Precision
is reduced

Andrew Thangaraj (IIT Madras) Continuous random variables 4 / 27

Bernoulli trials, n = 100, p = 0.5

Andrew Thangaraj (IIT Madras) Continuous random variables 5 / 27

Bernoulli trials, n = 100, p = 0.5

Calculations with the PMF is not very easy even in this case
Working with intervals and histograms can be much simpler

Andrew Thangaraj (IIT Madras) Continuous random variables 5 / 27

 Subsection 2

Cumulative distribution function

Andrew Thangaraj (IIT Madras) Continuous random variables 6 / 27

CDF of a random variable

Definition (CDF of a random variable)

The Cumulative Distribution Function (CDF) of a random variable X ,
denoted FX (x ), is a function from R to [0, 1], defined as

FX (x ) = P(X ≤ x ).

Andrew Thangaraj (IIT Madras) Continuous random variables 7 / 27

CDF of a random variable

Definition (CDF of a random variable)

The Cumulative Distribution Function (CDF) of a random variable X ,
denoted FX (x ), is a function from R to [0, 1], defined as

FX (x ) = P(X ≤ x ).

Properties
FX (b) − FX (a) = P(a < X ≤ b)
FX : non-decreasing function taking non-negative values
As x → −∞, FX goes to 0.
As x → ∞, FX goes to 1.

Andrew Thangaraj (IIT Madras) Continuous random variables 7 / 27

Example: Bernoulli random variable

1−p p
X ∼ { 0 , 1}


0

 x <0
FX (x ) = 1−p 0≤x <1

1≤x

1

Andrew Thangaraj (IIT Madras) Continuous random variables 8 / 27

Example: Throw a die

1/6 1/6 1/6 1/6 1/6 1/6

X ∼{1, 2, 3, 4, 5, 6}


0

 x <1

1/6 1≤x <2






2/6


 2≤x <3
FX (x ) = 3/6 3 ≤ x < 4

4/6 4 ≤ x < 5






5/6 5 ≤ x < 6





6≤x

1

Andrew Thangaraj (IIT Madras) Continuous random variables 9 / 27

CDF of a discrete random variable
p1 p2 p3 p4 p5
X ∼ {x1 , x2 , x3 , x4 , x5 }

Andrew Thangaraj (IIT Madras) Continuous random variables 10 / 27

CDF of a discrete random variable
p1 p2 p3 p4 p5
X ∼ {x1 , x2 , x3 , x4 , x5 }

Andrew Thangaraj (IIT Madras) Continuous random variables 10 / 27

Computing probability of intervals using CDF

X ∼ Uniform{1, 2, . . . , 100}


0

 x ≤0
FX (x ) = k/100 k ≤ x < k + 1, k = 1, 2, . . . , 99

x ≥ 100

1

Andrew Thangaraj (IIT Madras) Continuous random variables 11 / 27

Computing probability of intervals using CDF

X ∼ Uniform{1, 2, . . . , 100}


0

 x ≤0
FX (x ) = k/100 k ≤ x < k + 1, k = 1, 2, . . . , 99

x ≥ 100

1

P(3 < X ≤ 10) = FX (10) − FX (3) = 7/100

P(3.2 < X ≤ 10.6) = FX (10.6) − FX (3.2) = 7/100
P(X ≤ 17) = FX (17) = 17/100
P(X ≤ 17.3) = FX (17.3) = 17/100
P(X > 87) = 1 − FX (87) = 13/100
P(X > 87.4) = 1 − FX (87.4) = 13/100

Andrew Thangaraj (IIT Madras) Continuous random variables 11 / 27

Large alphabet: Binomial(n, 0.6)

Andrew Thangaraj (IIT Madras) Continuous random variables 12 / 27

Large alphabet: Meteorite data

Histogram of weights CDF of weight distribution

Andrew Thangaraj (IIT Madras) Continuous random variables 13 / 27

Cumulative Distribution Functions

Definition (CDF)
A function F : R → [0, 1] is said to be a Cumulative Distribution Function
(CDF) if
1 F is a non-decreasing function taking values between 0 and 1.

2 As x → −∞, F goes to 0.
x
3 As x → ∞, F goes to 1.
x
4 Technical: F is continuous from the right.

Andrew Thangaraj (IIT Madras) Continuous random variables 14 / 27

Cumulative Distribution Functions

Definition (CDF)
A function F : R → [0, 1] is said to be a Cumulative Distribution Function
(CDF) if
1 F is a non-decreasing function taking values between 0 and 1.

2 As x → −∞, F goes to 0.
x
3 As x → ∞, F goes to 1.
x
4 Technical: F is continuous from the right.

Definition motivated by CDF of a random variable defined earlier

A general CDF need not be like a CDF of a discrete random variable
I No need for a step-like structure
I Can be smooth and continuous

Andrew Thangaraj (IIT Madras) Continuous random variables 14 / 27

Examples of valid CDFs

Andrew Thangaraj (IIT Madras) Continuous random variables 15 / 27

Examples of valid CDFs

Continuous CDFs appear to be close approximations to CDFs of

discrete random variables, particularly when alphabet grows.
Andrew Thangaraj (IIT Madras) Continuous random variables 15 / 27
Probability of intervals using continuous CDF

X ∼ Uniform{1, 2, . . . , 100}


0
 x ≤0 


0
 x ≤0
k/100 k ≤x <k +1
 
FX (x ) = F (x ) = x /100 0 ≤ x ≤ 100
 k = 1, . . . , 99 
x ≥ 100

 
1

1 x ≥ 100



Andrew Thangaraj (IIT Madras) Continuous random variables 16 / 27

Probability of intervals using continuous CDF

X ∼ Uniform{1, 2, . . . , 100}


0
 x ≤0 


0
 x ≤0
k/100 k ≤x <k +1
 
FX (x ) = F (x ) = x /100 0 ≤ x ≤ 100
 k = 1, . . . , 99 
x ≥ 100

 
1

1 x ≥ 100



P(3 < X ≤ 10) = FX (10) − FX (3) = 7/100 = F (10) − F (3)

P(3.2 < X ≤ 10.6) = FX (10.6) − FX (3.2) = 7/100 ≈
F (10.6) − F (3.2) = 7.4/100
P(X ≤ 17.3) = FX (17.3) = 17/100 ≈ F (17.3) = 17.3/100

Andrew Thangaraj (IIT Madras) Continuous random variables 16 / 27

Probability of intervals using continuous CDF

X ∼ Uniform{1, 2, . . . , 100}


0
 x ≤0 


0
 x ≤0
k/100 k ≤x <k +1
 
FX (x ) = F (x ) = x /100 0 ≤ x ≤ 100
 k = 1, . . . , 99 
x ≥ 100

 
1

1 x ≥ 100



P(3 < X ≤ 10) = FX (10) − FX (3) = 7/100 = F (10) − F (3)

P(3.2 < X ≤ 10.6) = FX (10.6) − FX (3.2) = 7/100 ≈
F (10.6) − F (3.2) = 7.4/100
P(X ≤ 17.3) = FX (17.3) = 17/100 ≈ F (17.3) = 17.3/100
Simpler continuous CDF approximates prob of intervals!

Andrew Thangaraj (IIT Madras) Continuous random variables 16 / 27

Binomial using continuous CDF

X ∼ Binomial(100, 0.6)

k 1
!
X 100 F (x ) =
FX (k) = (0.6)j (0.4)n−j 
−1.65451(x −60)

j=0
j 1 + exp √
24

Andrew Thangaraj (IIT Madras) Continuous random variables 17 / 27

Binomial using continuous CDF

X ∼ Binomial(100, 0.6)

k 1
!
X 100 F (x ) =
FX (k) = (0.6)j (0.4)n−j 
−1.65451(x −60)

j=0
j 1 + exp √
24

P(40 < X ≤ 50) = 0.0271, F (50) − F (40) = 0.0318

P(50 < X ≤ 60) = 0.5108, F (60) − F (50) = 0.4670
P(60 < X ≤ 70) = 0.4473, F (70) − F (60) = 0.4670
P(70 < X ≤ 80) = 0.0148, F (80) − F (70) = 0.0318

Andrew Thangaraj (IIT Madras) Continuous random variables 17 / 27

Binomial using continuous CDF

X ∼ Binomial(100, 0.6)

k 1
!
X 100 F (x ) =
FX (k) = (0.6)j (0.4)n−j 
−1.65451(x −60)

j=0
j 1 + exp √
24

P(40 < X ≤ 50) = 0.0271, F (50) − F (40) = 0.0318

P(50 < X ≤ 60) = 0.5108, F (60) − F (50) = 0.4670
P(60 < X ≤ 70) = 0.4473, F (70) − F (60) = 0.4670
P(70 < X ≤ 80) = 0.0148, F (80) − F (70) = 0.0318
Better approximations are possible, partcularly as n grows!

Andrew Thangaraj (IIT Madras) Continuous random variables 17 / 27

 Subsection 3

General random variables and Continuous random variables

Andrew Thangaraj (IIT Madras) Continuous random variables 18 / 64

CDFs and random variables

Theorem (Random variable with CDF F (x ))

Given a valid CDF F (x ), there exists a random variable X taking values in
R such that
P(X ≤ x ) = F (x ).

Andrew Thangaraj (IIT Madras) Continuous random variables 19 / 64

CDFs and random variables

Theorem (Random variable with CDF F (x ))

Given a valid CDF F (x ), there exists a random variable X taking values in
R such that
P(X ≤ x ) = F (x ).

Properties
P(a < X ≤ b) = F (b) − F (a)

Andrew Thangaraj (IIT Madras) Continuous random variables 19 / 64

CDFs and random variables

Theorem (Random variable with CDF F (x ))

Given a valid CDF F (x ), there exists a random variable X taking values in
R such that
P(X ≤ x ) = F (x ).

Properties
P(a < X ≤ b) = F (b) − F (a)
If F (x ) rises from F1 to F2 at x1 , P(X = x1 ) = F2 − F1

Andrew Thangaraj (IIT Madras) Continuous random variables 19 / 64

CDFs and random variables

Theorem (Random variable with CDF F (x ))

Given a valid CDF F (x ), there exists a random variable X taking values in
R such that
P(X ≤ x ) = F (x ).

Properties
P(a < X ≤ b) = F (b) − F (a)
If F (x ) rises from F1 to F2 at x1 , P(X = x1 ) = F2 − F1
If F (x ) is continuous at x0 , P(X = x0 ) = 0 (non-intuitive!)

Andrew Thangaraj (IIT Madras) Continuous random variables 19 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002
I Value with finite precision taken with positive probability

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002
I Value with finite precision taken with positive probability
P(1.9999999 < X ≤ 2.0000001) = 0.00000002

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002
I Value with finite precision taken with positive probability
P(1.9999999 < X ≤ 2.0000001) = 0.00000002
I As precision increases, probability decreases

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002
I Value with finite precision taken with positive probability
P(1.9999999 < X ≤ 2.0000001) = 0.00000002
I As precision increases, probability decreases
P(X = 2.00000 · · · ) = 0

Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64

Example: Random variable X with CDF F (x )


0 x <0


F (x ) = 0.5 + 0.1x 0≤x ≤5


1 x >5

P(X = 0) = 0.5
P(1.99 < X ≤ 2.01) = F (2.01) − F (1.99) = 0.002
I Value with finite precision taken with positive probability
P(1.9999999 < X ≤ 2.0000001) = 0.00000002
I As precision increases, probability decreases
P(X = 2.00000 · · · ) = 0
I Values with infinte precision cannot be taken, when F (x ) is continuous
at that point
Andrew Thangaraj (IIT Madras) Continuous random variables 20 / 64
Continuous random variable
Definition (Continuous random variable)
A random variable X with CDF FX (x ) is said to be a continuous random
variable if FX (x ) is continuous at every x .

Andrew Thangaraj (IIT Madras) Continuous random variables 21 / 64

Continuous random variable
Definition (Continuous random variable)
A random variable X with CDF FX (x ) is said to be a continuous random
variable if FX (x ) is continuous at every x .

CDF has no jumps or steps

So, P(X = x ) = 0 for all x
Probability of X falling in an interval will be nonzero

P(a < X ≤ b) = F (b) − F (a)

Since P(X = a) = 0 and P(X = b) = 0, we have

P(a ≤ X ≤ b) = P(a < X ≤ b) = P(a ≤ X < b) = P(a < X < b)

Andrew Thangaraj (IIT Madras) Continuous random variables 21 / 64

Examples

Andrew Thangaraj (IIT Madras) Continuous random variables 22 / 64

Some scenarios for continuous models

Throw a dart onto a circular board - distance of the point of impact

from the center of the board.
Weight of a metoerite hitting earth
Weight of a human being, height of a human being
Speed of a delivery in cricket
Price of a stock
Many discrete random variables are well-approximated by continuous
random variables that are much simpler to describe

Andrew Thangaraj (IIT Madras) Continuous random variables 23 / 64

Problem
Consider a random variable X with CDF



0 x < −5

0.2 −5 ≤ x < 0

FX (x ) =


0.2 + 0.2x 0≤x <4

1 x ≥ 4.



Find P(X < −3), P(−3 < X < −1), P(−1 < X < 1), P(X ≤ −3),
P(0 ≤ X < 3). Is there an x0 for which P(X = x0 ) > 0? Is X a continuous
random variable?

Andrew Thangaraj (IIT Madras) Continuous random variables 24 / 64

Working

Andrew Thangaraj (IIT Madras) Continuous random variables 25 / 64

Problem
Consider a random variable X with CDF



 0 x < −5

0.04x + 0.2 −5 ≤ x < 0

FX (x ) =


 0.2 + 0.2x 0≤x <4

1 x ≥ 4.



Find P(X < −3), P(−3 < X < −1), P(−1 < X < 1), P(X ≤ −3),
P(0 ≤ X < 3). Is there an x0 for which P(X = x0 ) > 0? Is X a continuous
random variable?

Andrew Thangaraj (IIT Madras) Continuous random variables 26 / 64

Working

Andrew Thangaraj (IIT Madras) Continuous random variables 27 / 64

 Subsection 4

Probability density function and common continuous distributions

Andrew Thangaraj (IIT Madras) Continuous random variables 28 / 64

Refresher on integration
Indefinite integral of a function f (x )
dF (x )
I A function F (x ) such that = f (x )
dx
R
I Denoted as F (x ) = f (x )dx

Definite integral of a function f (x )

I Suppose F (x ) is the indefinite integral of f (x )
I Definite integral of f (x ) from a to b is defined as
Z b
f (x )dx = F (b) − F (a)
a

I Definite integral equals the area under the curve f (x ) from a to b

Tables of integrals: [Link]

Andrew Thangaraj (IIT Madras) Continuous random variables 29 / 64

Probability density function (PDF)
Definition (PDF)
A continuous random variable X with CDF FX (x ) is said to have a PDF
fX (x ) if, for all x0 , Z x0
FX (x0 ) = fX (x )dx .
−∞

Andrew Thangaraj (IIT Madras) Continuous random variables 30 / 64

Probability density function (PDF)
Definition (PDF)
A continuous random variable X with CDF FX (x ) is said to have a PDF
fX (x ) if, for all x0 , Z x0
FX (x0 ) = fX (x )dx .
−∞

CDF is the integral of the PDF

I Derivative of the CDF (wherever it exists) is usually taken as the PDF

Andrew Thangaraj (IIT Madras) Continuous random variables 30 / 64

Probability density function (PDF)
Definition (PDF)
A continuous random variable X with CDF FX (x ) is said to have a PDF
fX (x ) if, for all x0 , Z x0
FX (x0 ) = fX (x )dx .
−∞

CDF is the integral of the PDF

I Derivative of the CDF (wherever it exists) is usually taken as the PDF

Why PDF?
I Value of PDF around fX (x0 ) is related to X taking a value around x0
F Higher the PDF, higher the chance that X lies there
F Contrast with value of CDF at x0 , FX (x0 )
F PDF is much easier in probability computations

Andrew Thangaraj (IIT Madras) Continuous random variables 30 / 64

Examples: Uniform distribution

Andrew Thangaraj (IIT Madras) Continuous random variables 31 / 64

Examples: Exponential and normal distribution

Andrew Thangaraj (IIT Madras) Continuous random variables 32 / 64

Properties of PDF
Definition (Density function)
A function f : R → R is said to be a density function if
1 f (x ) ≥ 0
R∞
−∞ f (x )dx = 1
2

3 f (x ) is piecewise continuous

Andrew Thangaraj (IIT Madras) Continuous random variables 33 / 64

Properties of PDF
Definition (Density function)
A function f : R → R is said to be a density function if
1 f (x ) ≥ 0
R∞
−∞ f (x )dx = 1
2

3 f (x ) is piecewise continuous

Given a density function f , there is a continuous random variable X

with PDF as f
Support of the random variable X with PDF fX is
supp(X ) = {x : fX (x ) > 0}
I supp(X ) contains intervals in which X can fall with positive probability
I Remember: P(X = x ) = 0 for a continuous random variable

For a random variable X with PDF fX , an event A isR

a subset of the
real line and its probability is computed as P(A) = A f (x )dx
Andrew Thangaraj (IIT Madras) Continuous random variables 33 / 64
Problem
Consider the function
(
3x 2 0<x <1
f (x ) =
0 otherwise.

Show that f is a density function. Consider a random variable X with

density f . Find P(X = 1/5), P(X = 2/5), P(X ∈ [1/5 − , 1/5 + ]),
P(X ∈ [2/5 − , 2/5 + ]).

Andrew Thangaraj (IIT Madras) Continuous random variables 34 / 64

Problem
Consider a random variable X with density
(
2x 0<x <1
fX (x ) =
0 otherwise.

Find P(X ∈ [0.1, 0.3]), P(X ∈ (0.1, 0.03]), P(X ∈ [0.1, 0.03)),
P(X ∈ (0.1, 0.03)).

Andrew Thangaraj (IIT Madras) Continuous random variables 35 / 64

Problem
Consider the function



 k 0 ≤ x < 1/4

2k 1/4 ≤ x < 3/4

f (x ) =


 3k 3/4 ≤ x < 1

0 otherwise.



Find k such that f (x ) is a valid density function.

Andrew Thangaraj (IIT Madras) Continuous random variables 36 / 64

Common distributions: Uniform

X ∼ Uniform[a, b]
PDF
(
1
b−a a<x <b
fX (x ) =
0 otherwise

CDF

0

 x ≤a
FX (x ) = x −a
a<x <b
 b−a
x ≥b

1

Andrew Thangaraj (IIT Madras) Continuous random variables 37 / 64

Probability computations with uniform distribution
Suppose X ∼ Uniform[−10, 10]. Find P(−3 ≤ X ≤ 2), P(5 < |X | < 7),
P(1 −  < X < 1 + ), P(9 −  < X < 9 + ), P(X > 7|X > 3).

Andrew Thangaraj (IIT Madras) Continuous random variables 38 / 64

Common distributions: Exponential

X ∼ Exp(λ)
PDF
(
λ exp(−λx ) x >0
fX (x ) =
0 otherwise

CDF
(
0 x ≤0
FX (x ) =
1 − exp(−λx ) x >0

Andrew Thangaraj (IIT Madras) Continuous random variables 39 / 64

Probability computations with exponential distribution
Suppose X ∼ Exp(2). Find P(5 < X < 7), P(1 −  < X < 1 + ),
P(9 −  < X < 9 + ), P(X > 4), P(X > 7|X > 3).

Andrew Thangaraj (IIT Madras) Continuous random variables 40 / 64

Common distributions: Normal

X ∼ Normal(µ, σ 2 )
PDF
!
1 (x − µ)2
fX (x ) = √ exp −
σ 2π 2σ 2

CDF
Z x
FX (x ) = fX (u)du
−∞

Standard normal: Normal(0, 1)

Andrew Thangaraj (IIT Madras) Continuous random variables 41 / 64

Probability computations with normal distribution
CDF of X ∼ Normal(µ, σ 2 ) does not have a closed form expression
Standardization: If X ∼ Normal(µ, σ 2 ), then
(X − µ)/σ ∼ Normal(0, 1)
I Z ∼ Normal(0, 1), PDF: fZ (z) = √1 exp(−z 2 /2)
2π
Z z
1
CDF: FZ (z) = √ exp(−u 2 /2)du
−∞ 2π

I Normal table: Tabulation of above function FZ (z)

F Available on most computing systems

How to compute probabilities for a normal distribution?

I Convert probability computation to that of a standard normal
I Use normal tables or computing systems

Andrew Thangaraj (IIT Madras) Continuous random variables 42 / 64

Problem
Suppose X ∼ Normal(2, 5). Find P(X < 5), P(X < 10), P(X < −5),
P(X < −10), P(X > 5), P(X > 10).

Andrew Thangaraj (IIT Madras) Continuous random variables 43 / 64

Problem
Suppose X ∼ Normal(3, 1). Find P(5 < X < 7), P(−5 < X < 5),
P(1 −  < X < 1 + ), P(9 −  < X < 9 + ), P(X > 4), P(X > 7|X > 3).

Andrew Thangaraj (IIT Madras) Continuous random variables 44 / 64