Scribd
Upload
9 views21 pages

Probability Lec 05

Lecture note

Uploaded by

snsfaojia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
9 views21 pages

Probability Lec 05

Lecture note

Uploaded by

snsfaojia
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 21

Axioms of Probability

Lecture 5

Jahida Gulshan

Institute of Statistical Research and Training, University of Dhaka

3 September, 2023
Probability for Discrete Sample Spaces

Sample space Ω is said to be discrete if it is countable.

The probability measure P can be defined by first assigning


probabilities to outcomes {ω}, such that: P(ω) ≥ 0, for all ω ∈ Ω,
and Σω∈Ω P(ω) = 1.

The probability of any event A (by the additivity axiom) is simply


P(A) = Σω∈A P(ω)
Probability for Continuous Sample Space

A sample space is continuous, if Ω is uncountably infinite.

For continuous Ω, we cannot, in general, define the probability


measure P by first assigning probabilities to outcomes.

In this case the probability of each single outcome event is zero.

Probability of an event is defined as an area in case of a continuous


sample space.
Probability Measure

The modern definition of probability, based on set and measure


theory is known as axiomatic definition of probability. It was not
defined as a ratio or as the limit of a ratio, but simply as a function
satisfying a set of axioms (Kolmogorov axioms).
The Kolmogorov axioms are the foundations of probability theory
introduced by Russian mathematician Andrey Kolmogorov in 1933.
These axioms remain central and have direct contributions to
mathematics, the physical sciences, and real-world probability cases.
Probability Measure

For developing the probability model for an experiment in a


Probability measure approach, we need to consider all possible
outcomes of an experiment.

The aggregate of all possible outcomes is defined as a sample space


(Ω). The space Ω is associated with a class A of subsets A.
We require that A be a σ-algebra of subsets of Ω i.e. A be
non-empty and satisfy
1. If A1 , A2 , · · · ∈ A, then ∪∞
i=1 Ai ∈ A.
2. If A ∈ Ω then Ω − A ∈ Ω.
Example of σ-algebra

For example, let Ω = all possible outcomes of tossing a coin.


Ω = {T , H}.
Then A = {ϕ, {T }, {H}, {H, T }}.

Here A consists of 4 subsets of Ω.

Again, let Ω = all possible outcomes of tossing a coin twice.


Then, Ω = {TT , TH, HT , HH}. Then

A = {ϕ, {TT }, {TH}, {HT }, {HH}, {TT , TH}, , {TT , HT },


{ TT, HH }, { TH, HT }, { TH, HH }, { HT, HH },
{ TT, TH, HT }, { TT, TH, HH }, { TT, HT, HH },
{ TH, HT, HH }, { TT, TH, HT, HH } }.
Here A consists of 16 subsets of Ω.
Measurable Space

The combination of Ω and A of the subsets of Ω is known as


measurable space {Ω, A}.
The term measurable indicates that we have a structure upon
which we define a measure or a probability measure.
Measure

µ(A) is a measure on {Ω, A} if it satisfies the conditions:


1. For each A, µ(A) is a real number
2. For each A, µ(A) ≥ 0
′
3. µ(∪∞ ∞
i=1 Ai ) = Σi=1 µ(Ai ) when Ai ∩ Aj = ϕ i.e. Ai s are disjoint.
Thus, a measure is an assignment of a number µ(A) to each set in
a class A.
For example, if Ω is a sample space of an experiment, certain
subsets of Ω will be called ‘events’ and µ(A) is an assigned number
to an event A which we call a measure.
Probability Measure

If we add one additional restriction, (Ω) = 1, to the definition of a


measure µ(A), we call it a probability measure. We denote
probability measures usually by P(A).

The axiomatic definition of probability is thus as follows.


Axiomatic Definition of Probability

A Probability measure ( or law or function) is an assignment of


probabilities to events such that the following three axioms are
satisfied:
1. P(A) ≥ 0, for all A (nonnegativity)
2. P(Ω) = 1 (normalization)
3. If A and B are disjoint (A ∩ B = ϕ), then
P(A ∪ B) = P(A) + P(B) (additivity).
More generally,
If the sample space has an infinite number of points and
A1 , A2 , · · · are disjoint events, then P(∪∞ ∞
i=1 Ai ) = Σi=1 P(Ai )
Motivation of Axioms of Probability

These axioms are motivated by the classical and frequency


definitions of probability.

Axiom 1 states that probabilities are non-negative real numbers,

Axiom 2 states that the sample space as a whole is assigned a


probability of 1, and thus probability is a number between 0 and 1,
inclusive.

Axiom 3 states that, for any sequence of mutually exclusive events,


the probability of at least one of these events is just the sum of
their respective probabilities.
When sample space is an uncountable infinite set, P(A) is defined
only for a class of measurable events.
However, all events of any practical interest are measurable.
The axiomatic definition of probability is, in fact, related to the
measure theory, where probability is defined as a particular type of
measure.
Probability Space

The triplet (Ω, A, P(.)) is called probability space, where Ω is the


sample space, A is a collection of events or set of subsets of Ω
(assumed to be a σ-algebra), and P(.) is a probability measure.
Some Basic Rules of Probability

Sometimes, we can write the event of interest in the form of unions,


complement or intersection of events. There are probability rules
that can simplify the computation of probabilities in such cases.
The Rule of Complement

Let A be any event defined on the sample space Ω and Ac be its


complement. Then probability of the complement Ac is given by

P(Ac ) = 1 − P(A)
This is known as rule of complement or complementary rule of
probability.
Rule of Complement: P(Ac ) = 1 − P(A)

Proof:
A and Ac are mutually exclusive, and A ∪ Ac = Ω. Since A and Ac
are mutually exclusive, so by Axiom 3,

P(A ∪ Ac ) = P(A) + P(Ac )


or,
P(Ω) = P(A) + P(Ac ).
By Axiom 2, P(Ω) = 1. So,

1 = P(A) + P(Ac )

and thus
P(Ac ) = 1 − P(A).
Probability of empty set: P(ϕ) = 0

Proof:
Φ = Ωc .
So, using the rule of complement,

P(ϕ) = 1 − P(Ω) = 1 − 1 = 0

.
If A ⊂ B, then P(A) ≤ P(B).

Proof:
Since A ⊂ B, we may decompose B into two mutually exclusive
events, A and B ∩ Ac , such that

B = A ∪ (B ∩ Ac ).

Then by axiom 3,

P(B) = P(A) + P(B ∩ Ac ).

Since P(B ∩ Ac ) ≥ 0, we have P(A) ≤ P(B).


For any event A, P(A) ≤ 1.

Proof:
A ⊂ Ω and P(Ω) = 1. Since A ⊂ B, implies that ,P(A) ≤ P(B),
we can write,
P(A) ≤ P(Ω)
or,
P(A) ≤ 1.
If A is an event defined on the sample space Ω, then
0 ≤ P(A) ≤ 1.

Proof. We see that


Φ ⊂ A ⊂ Ω.
We also known that, if A ⊂ B, then

P(A) ≤ P(B).

Therefore,
P(ϕ) ≤ P(A) ≤ P(Ω)
which leads to
0 ≤ P(A) ≤ 1.
If A is an event defined on the sample space Ω, then
0 ≤ P(A) ≤ 1.

Proof
We see that
Φ ⊂ A ⊂ Ω.
We also known that, if A ⊂ B, then

P(A) ≤ P(B).

Therefore,
P(ϕ) ≤ P(A) ≤ P(Ω)
which leads to
0 ≤ P(A) ≤ 1.

You might also like