Scribd
Upload
15 views

Lecture 2

Uploaded by

Anika Gupta
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
15 views

Lecture 2

Uploaded by

Anika Gupta
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

Axiomatic Approach to probability is an improvement upon all the above approaches

to probability. It is called the unifying theory of probability as the axioms stated in


this approach apply to both classical probability and empirical probability.

For defining probability through axiomatic approach, we need an understanding of


the following concepts:

i) 𝑆𝑖𝑔𝑚𝑎 𝐴𝑙𝑔𝑒𝑏𝑟𝑎: Let 𝑆 be the collection of all possible outcomes (called Sample
Space) of an experiment. Let ℬ ⊆ 𝑃(𝑆), the power set of 𝑆 .Which means that ℬ is a
collection of subsets of 𝑆. Then, ℬ is called a sigma algebra if it satisfies the following:
(a) Øϵℬ
(b) 𝐴ϵℬ⇒ 𝐴𝑐 𝜖ℬ (ℬ is closed under complementation)
(c) 𝐴1 , 𝐴2 , 𝐴3 , …ϵℬ⇒⋃∞
𝑖=1 𝐴𝑖 𝜖ℬ (ℬ is closed under countable union).

𝐸𝑥𝑎𝑚𝑝𝑙𝑒𝑠:
a) Consider a subset 𝐴 ∈ 𝑆, where 𝑆 is the Sample Space (or collection of all possible
outcomes) of some experiment performed. Let ℬ={Ø, 𝐴, 𝐴𝑐 , 𝑆}. Then ℬ is a sigma
algebra.

b) Let ℬ= 𝑃(𝑆), where 𝑆 is the Sample Space of some experiment performed. Then ℬ is
a sigma algebra.

ii) 𝑆𝑒𝑡 𝐹𝑢𝑛𝑐𝑡𝑖𝑜𝑛: A set function is a function whose domain is a family/collection of


subsets of a given set and codomain is the set of real numbers.

𝐸𝑥𝑎𝑚𝑝𝑙𝑒𝑠:

a) 𝐷𝑒𝑓𝑖𝑛𝑒 𝑓: ℬ→ ℝ as 𝑓([𝑎, 𝑏]) = 𝑏 − 𝑎 = 𝑙(𝑎, 𝑏), where ℬ is the collection of all


intervals of finite length. Function 𝑓 which is defined as the length of an interval
is a set function as it is defined on ℬ, a collection of sets (subsets of ℝ).

b) 𝐷𝑒𝑓𝑖𝑛𝑒 𝑔: ℳ → ℝ as 𝑔({𝑎1 , 𝑎2 , 𝑎3 , … , 𝑎𝑛 }) = 𝑛 =
𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑒𝑙𝑒𝑚𝑒𝑛𝑡𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑠𝑒𝑡 {𝑎1 , 𝑎2 , 𝑎3 , … , 𝑎𝑛 }, where ℳ is the collection of
all finite sets. Function 𝑔 which is defined as the cardinality of a set is a set
function as it defined on a collection of sets.

2.2.4 𝐴𝑥𝑖𝑜𝑚𝑎𝑡𝑖𝑐 𝑃𝑟𝑜𝑏𝑎𝑏𝑖𝑙𝑖𝑡𝑦

Let 𝑆 be the collection of all possible outcomes / Sample Space of an experiment. Let ℬ
be a sigma algebra of subsets of 𝑆. Define a set function 𝑃: ℬ → ℝ which takes 𝐴 ∈ ℬ →
𝑃(𝐴) ∈ ℝ. This set function 𝑃 is called a Probability function if it satisfies the following
three axioms given by Russian Mathematician Kolmogorov.

(a) 𝑁𝑜𝑛 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑖𝑡𝑦: 𝑃(𝐴) ≥ 0 ∀ 𝐴𝜖ℬ.

1
(b) 𝐶𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦: 𝑃(𝑆) = 1
(c) 𝐷𝑒𝑛𝑢𝑚𝑒𝑟𝑎𝑏𝑙𝑒/ 𝐶𝑜𝑢𝑛𝑡𝑎𝑏𝑙𝑒 𝐴𝑑𝑑𝑖𝑡𝑖𝑣𝑖𝑡𝑦: Let 𝐴1 , 𝐴2 , 𝐴3 , … 𝜖ℬ such that 𝐴𝑖 ⋂ 𝐴𝑗 =Ø.
Then, 𝑃(⋃∞ ∞
𝑖=1 𝐴𝑖 ) = ∑𝑖=1 𝑃( 𝐴𝑖 ).

In words, the first axiom states that probability of every event is non-negative. The
second axiom states that probability of sample space/universal set is one. The third
axiom says that the probability of the union of mutually exclusive members of ℬ is
the sum of their individual probabilities. 𝑆 is called the 𝑠𝑎𝑚𝑝𝑙𝑒 𝑠𝑝𝑎𝑐𝑒 of the
experiment and ℬ is called the 𝑒𝑣𝑒𝑛𝑡 𝑠𝑝𝑎𝑐𝑒. The members of ℬ are called 𝑒𝑣𝑒𝑛𝑡𝑠.

Let us discuss a few basic properties of Probability Function:

𝑃𝑟𝑜𝑝𝑒𝑟𝑡𝑖𝑒𝑠:

1) Probability of complement of an event is obtained in terms of the probability


of the original event.

𝑃(𝐴𝑐 )=1 − 𝑃(𝐴) where 𝐴𝑐 is the complement of event 𝐴 i.e., collection of all
those outcomes in 𝑆 which are not in 𝐴.(See Fig 2.1)

Fig 2.1

As 𝐴 ∪ 𝐴𝑐 = 𝑆 ⇒ 𝑃(𝐴 ∪ 𝐴𝑐 ) = 𝑃(𝑆) = 1. Therefore, using 𝐴 ∩ 𝐴𝑐 = ∅, we get 𝑃(𝐴) + 𝑃(𝐴𝑐 ) =


1 ⇒ 𝑃(𝐴𝑐 ) = 1 − 𝑃(𝐴).

2) The probability of impossible event is zero.


𝑃(∅) = 0

2
Ø=𝑆 𝑐 ⇒ 𝑃(Ø) = 𝑃(𝑆 𝑐 ) = 1 − 𝑃(𝑆) = 1 − 1 = 0.

3) Probability function is a monotone function. It preserves order relation.


𝐴 ⊆ 𝐵 ⇒ 𝑃(𝐴) ≤ 𝑃(𝐵).(Fig 2.2)

Fig 2.2

4) The probability of difference of two events can be obtained in terms of the


difference of individual probability of one event and the probability of
intersection of two events. (Fig 2.3)

(a) 𝑃(𝐴~𝐵) = 𝑃(𝐴) − 𝑃(𝐴 ∩ 𝐵)


(b) 𝑃(𝐵~𝐴) = 𝑃(𝐵) − 𝑃(𝐴 ∩ 𝐵) ∀𝐴, 𝐵𝜖ℬ

Fig 2.3

(a) We can write 𝐴 = ( 𝐴~𝐵)∪ ( 𝐴 ∩ 𝐵).


Therefore, 𝑃(𝐴) = 𝑃[(𝐴~𝐵)∪ ( 𝐴 ∩ 𝐵)] = 𝑃(𝐴~𝐵) + 𝑃(𝐴 ∩ 𝐵) 𝑎𝑠 𝐴~𝐵 𝑎𝑛𝑑 𝐴 ∩
𝐵 𝑎𝑟𝑒 𝑚𝑢𝑡𝑢𝑎𝑙𝑙𝑦 𝑒𝑥𝑐𝑙𝑢𝑠𝑖𝑣𝑒.
Therefore, 𝑃(𝐴~𝐵) = 𝑃(𝐴) − 𝑃(𝐴 ∩ 𝐵).

(b) We can similarly write 𝐵 = ( 𝐵~𝐴)∪ ( 𝐴 ∩ 𝐵).

3
Therefore, 𝑃(𝐵) = 𝑃[(𝐵~𝐴)∪ ( 𝐴 ∩ 𝐵)] = 𝑃(𝐵~𝐴) + 𝑃(𝐴 ∩ 𝐵) 𝑎𝑠 𝐵~𝐴 𝑎𝑛𝑑 𝐴 ∩
𝐵 𝑎𝑟𝑒 𝑚𝑢𝑡𝑢𝑎𝑙𝑙𝑦 𝑒𝑥𝑐𝑙𝑢𝑠𝑖𝑣𝑒.
Therefore, 𝑃(𝐵~𝐴) = 𝑃(𝐵) − 𝑃(𝐴 ∩ 𝐵).

5) Probability of an event always lies between zero and one. Zero represents
impossibility while one represents certainty. Therefore, 𝑃: ℬ → [0,1].

0 ≤ 𝑃(𝐴) ≤ 1 ∀𝐴𝜖ℬ

∅ ⊆ 𝐴 ⊆ 𝑆⇒𝑃(∅) ≤ 𝑃(𝐴) ≤ 𝑃(𝑆) ⇒ 0 ≤ 𝑃(𝐴) ≤ 1.

You might also like