ADVANCED PROBABILITY AND MEASURE THEORY

SECTION-A

CHAPTER ONE: SETS AND CLASSES OF EVENT

Sets: A sets is the collection of well-defined distinct objects. The word well-defined refers
to a specific property which makes it easy to identify whether the given object belongs to the
set or not. The word „distinct‟ means that the objects of a set must be all different.

For example:

1. The collection of children in class VII whose weight exceeds 35 kg represents a set.
2. All natural numbers.
3. The vowels of the English alphabet: a, e, i, o, and u.
4. Books in a library.

Cartesian products: If A and B are sets, the Cartesian product of A and B is denoted by
A and defined as:

A *( ) ( ) ( )+

The following points are worth special attention: The Cartesian product of two sets is a set,
and the elements of that set are ordered pairs. In each ordered pair, the first component is an
element of A, and the second component is an element of B.

Example: If A= {1, 2}; and B= {3, 4). Then the Cartesian product of A and B is

A = {1, 2} * += {(1, 3), (1, 4), (2, 3), (2, 4)}

B {3, 4} * + *( )( )( )( )+

Relations: A relation from a set A to a set B is a subset of A . Any subset of A is

called a binary relation on A.

1
Example 1: Let A = {1, 2, 3, 4}. Which ordered pairs are in the relation R = {(a, b) | a < b}?
The relation is R= {(1, 2), (1, 3), (1, 4), (2, 3), (2, 4), (3, 4)}

Example 2: If A = {3, 4, 5} B = {2, 4, 6, 9, 15, 16, 25}, then relation R from A to B is

defined as „is a positive square root of‟. Here, relation is R = {(3, 9); (4, 16); (5, 25)}.

Example 3: If A = {2, 3, 4, 5} and B = {1, 3, 5} and R be the relation 'is less than' from A to
B, then R = {(2, 3), (2, 5), (3, 5), (4, 5)}

Domain and Range: If R be the relation from set A and B. The set of all first components of
the ordered pairs belongs to R is called the domain of R. Thus, Dom(R) = {a A: (a, b) R
for some b B}.

The set of all second components of the ordered pairs belonging to R is called the range of
R. Thus, range of R = {b B: (a, b) R for some a A}.

Note that:

• The domain of a relation from A to B is a subset of A.

• The range of a relation from A to B is a subset of B.

Example: If A = {2, 4, 6, 8) B = {5, 7, 1, 9}. Let R be the relation „is less than‟ from A to
B. Find Domain (R) and Range (R).

Solution: Under the relation (R), we have

R = {(4, 5); (4, 7); (4, 9); (6, 7); (6, 9), (8, 9) (2, 5) (2, 7) (2, 9)}

Therefore, Domain (R) = {2, 4, 6, 8} and Range (R) = {1, 5, 7, 9}.

Home Work:

1. If A = {2, 4, 6, 8) B = {5, 7, 1, 9}. Let R be the relation „is less than‟ from A to B.
Find Domain (R) and Range (R).

2. In the given ordered pair (4, 6); (8, 4); (4, 4); (9, 11); (6, 3); (3, 0); (2, 3) find the
following relations. Also, find the domain and range. (a) Is less than; (b) Is greater
than; and (c) Is equal to

2
 3. Let A = {2, 3, 4, 5} and B = {8, 9, 10, 11}. Let R be the relation „is factor of‟ from A
to B. (a) Write R in the roster form. Also, find Domain and Range of R. (b) Draw an
arrow diagram to represent the relation.

Different types of relations:

Reflexive: A relation R on a set A is called reflexive if (a, a) R for every element aA.

Example: Are the following relations on {1, 2, 3, 4} reflexive?

Relations Is R Reflexive?
• R = {(1, 1), (1, 2), (2, 3), (3, 3), (4, 4)} No
• R = {(1, 1), (2, 2), (2, 3), (3, 3), (4, 4)} Yes
• R = {(1, 1), (2, 2), (3, 3), (2,1), (3, 2)} No

Irreflexive: A relation on a set A is called irreflexive if (a, a) R for every element aA.

Symmetric: A relation R on a set A is called symmetric if (b, a) R whenever (a, b) R for

all a, bA.

Asymmetric: A relation R on a set A is called asymmetric if (a, b) R implies that (b, a) R

for all a, bA.

Example: Are the following relations on {1, 2, 3, 4} symmetric, antisymmetric, or

asymmetric?

Relations Is R Symmetric/Asymmetric?
• R1 = {(1, 1), (1, 2), (2, 1), (3, 3), (4, 4)} Yes
• R2={(1, 3), (1,4), (3,1), (2, 2), (4, 1)} Yes
• R3={(1, 1), (2, 2), (3, 3), (1, 3)} No
• R4={(1,2), (2, 1), (3, 4), (4, 3)} Yes
• R5={(1, 3), (3, 1), (3, 4), (4, 4)} No
• R6 = {(1, 3), (3, 2), (2, 1)} No
• R7={(1,1), (2, 2), (3, 3), (4, 4)} Yes

Transitive: A relation R on a set A is called transitive if whenever (a, b)R and (b, c)R,
then (a, c)R for a, b, cA.

Example: Are the following relations on {1, 2, 3, 4} transitive?

3
 Relations Is R Transitive?
• R = {(1, 1), (1, 2), (2, 2), (2, 3), (1, 3)} Yes
• R = {(1, 3), (3, 2), (2, 1)} No
• R = {(2, 4), (4, 3), (2, 3), (4, 1)} No

Equivalence: A relations R in a set A is an equivalence relations if

(i) R is Reflexive, that is for every a ( )

(ii) R is symmetric, that is (a, b) ( )
(iii) R is transitive, that is (a, b) ( ) ( )

Theorem: If a relations R is transitive, then its inverse relation R-1 is also transitive.

Proof: Let (a, b) and (b, c) belongs to R-1, then (b, a) ( ) Since R is
transitive, (c, a) also belongs to R; hence (a, c) . We have shown that (a, b) , (b,
c) implies (a, c) ; hence is transitive.

Theorem: Let R and Rʹ be the relations in a set A. Show that each of the following two
statements:

(i) If R is symmetric and Rʹ is symmetric, then R ʹ is also symmetric.

(ii) If R is reflexive and Rʹ is any relation, then R ʹ is reflexive.

Proof (i): If (a, b) R ʹ , then (a, b) belongs to R or ʹ , which are symmetric. Hence (b,
a) is also belongs to R or ʹ, then (b, a) ʹ ʹis symmetric.

Proof (ii): R is reflexive if and only if R contains the “diagonal line” D of A But
D ʹ implies that ʹ. Therefore, ʹ is reflexive.

Topology: Let X be a non-empty set. A class T of subsets of X is a topology on X if and

only if T satisfies the following properties:

1. The empty set ∅ and the space X are both sets in the topology.
2. The union of any collection of sets in τ is contained in τ.
3. The intersection of any finitely many sets in τ is also contained in τ.

4
If τ is a topology on X, then the pair (X, τ) is called a topological space, and the elements of τ
are called the open sets of (X, τ). When the topology is evident from the context, we speak
simply of the open sets of X. The elements of X are also called points.

Example: Consider the following classes of subsets of X = {a, b, c, d, e}. T1={X, ∅, {a}, {c,
d}, {a, c, d}, {b, c, d, e}} and T2={X, ∅, {a}, {c, d}, {a, c, d}, {b, c, d}. Prove that T1 is a
topology on X, but T2 is not topology.

Solution: Try yourself.

Discrete topology: Let D denote the class of all subsets of X. Where X be a non-empty set.
Then D satisfies the following axioms:

(i) X and ∅ belongs to D.

(ii) The union of any number of sets in D belongs to D.
(iii)The intersection of any two sets in D belongs to D.

This topology is called the discrete topology and X together with its discrete topology. i.e.,
the pair (X, D) is called a discrete topological space or simply a discrete space.

Examples: Consider the following classes of subsets of X= {a, b, c}.

D={X ∅, {a}, {b}, {c}, {a, b}, {b, c}, {c, a}}

Where, D contains all of the subset of the given set, and then D is a discrete topology. And
(X, D) together with called the discrete topological space.

Indiscrete topology: A topology on X must contain the sets X and ∅. The class A={X, ∅}
consists of X and ∅ alone, is itself a topology on X. It is called indiscrete topology with its
indiscrete topology. i.e, (X, A) is called an indiscrete topological space or simply an
indiscrete space.

Example: Consider the following classes of subsets X= {a, b, c} and A={X, ∅}. Where, A
contains only X and ∅. So A is an indiscrete topology, and the pair (X, A) is called the
indiscrete topological space.

Theorem: If T1 and T2 are topologies, then show that T1 T2 is also topology.

5
Solution: Let us consider two topologies T1 and T2 and X is an non-empty set. And X is also
a topology on X.

We can say that T1 T2 is a topology if it satisfies the following conditions:

(i) X and ∅ belongs to T1 T2.

(ii) The union of any number of sets in T1 T2 belongs to T1 T2.
(iii)The intersection of any two sets in T1 T2 belongs to T1 T2.

Since, X and ∅ each belongs to both T1 and T2. Hence X and ∅ each belongs to the
intersection T1 T2. i.e., T1 T2 satisfies (i).

Let, G and H be any two non-empty set. Then, if G, H , then in particular G,

H . But since T1 and T2 are topologies. G and G .

Accordingly, G .

In other words, satisfies (iii).

Similarly, satisfies (ii).

Thus, if and are topologies, then also topology.

Theorem: If T1 and T2 are topologies, then show that T1 T2 is not always topology.

Solution: We have to show that if T1 and T2 are topologies, then show that T1 T2 is not
always topology.

Let us consider,

T1={X, ∅, {a}} and T2=={X, ∅, {b}} is a topology on X= {a, b, c}.

We can say that T1 T2 is a topology if it satisfies the following conditions:

(i) X and ∅ belongs to T1 T2.

(ii) The union of any number of sets in T1 T2 belongs to T1 T2.
(iii)The intersection of any two sets in T1 T2 belongs to T1 T2.

Now, T1 T2={X, ∅, {a}} {X, ∅, {b}} ={X, ∅, {a}, {b}}

6
T1 T2 is not a topology on X since it violates (ii). That {a} T1 T2 and {b} T1 T2

But {a} {b} ={a, b} does not belongs to T1 T2.

Closed set: Let X be a topological space. A subset A of X is a closed set if and only if its
complement Ac is an open set.

Example: The class T={X, ∅, {a}, {c, d}, {a, c, d}, {b, c, d, e}} defines a topology on X=
{a, b, c, d, e}. The closed sunsets of X are

∅, X, {b, c, d, e}, {a, b, e}, {b, e}, {a}

CHAPTER TWO: MEASURE

Algebra/Field: Let X be non-empty set (space). A collection of subsets of X is

algebra/field on X if and only if it satisfies the following properties:

1. X,
2. For all A , then AC
3. For all A, B , then A B A

Thus algebra is classes of subset containing X that is closed under complementation and
closed under finite union.

For example:

1. The collection P(X) of all subsets of X is algebra/field.

2. The collection of ={X, } is an algebra.
3. Let X={1,2,3,4}, then the collection is
*X * +* ++, an algebra.

Example 1: If X= {a, b, c, d}, then the collection of subsets, *X * +* ++

and * * ++. Are and algebra? Justify your answer.

Example 2: Is the power set of X denoted by P (X) = 2n an algebra?

7
Example 3: Let X be the sample space and be an empty set. The collection of subset is
defined as: ={X, }. Is algebra? Justify your answer?

Theorem: Let be a collection of subsets of a set X with the following properties:

1. X
2. If A, B , then
Show that is an algebra/field.
Solution:

(i) X
(ii) If then X (by 2)
(iii) If Since (by (ii)
Since, ( ) . Thus .

Theorem:
(a) Show that if ( ) is an increasing sequence of algebras of subsets of a set X, and then
⋃ is an algebra of subsets of X.
(b) Show by an example that even if in (a) is a -algebra for every , the union still
may not be a -algebra.
Solution:
(a) Let A=⋃ . We have to show that A is an algebra.
(i) Since X , for all , so X
(ii) Let Then for some n. And also (Since is an algebra).
Thus .
(iii) Suppose . We shall show that .
Since * + is increasing. i.e., ⋃ there is some
such that . Thus Hence, .
(b) Let X=N, =the family of all subsets {1, 2, 3, …, n} and their complements. Clearly,
is a -algebra and . However, ⋃ is the family of all finite and co-
finite subsets of N, which is not a -algebra.

8
Theorem: Let X be an arbitrary in finite set. We say that a subset A of X is co-finite if its
complement is a finite subset of X. Let consists of all the finite and the co-finite subsets of
a set X.
(a) Show that is algebra of subsets of X.
(b) Show that is an -algebra if and only if X is a finite set

Solution:
(a)
(i) X . Since X is co-finite.
(ii) Let A . If A is finite then is co-finite, so . If A is co-finite then is
finite, so In both cases,
A
(iii) Let A . We shall show that If A and B are finite, then A B is
finite, so A B . Otherwise, assume that A is co-finite, then A B is co-finite, so
A B . In both cases
A .
(b) If X is finite then (X), which is -algebra. To show the reserve, i.e., if is a -
algebra then X is finite. So we can find an infinite sequence ( ) of distinct
elements of X such that X\* + is infinite. Let An={an}. Then for any
⋃ is neither finite nor co-finite. ⋃ . Thus is not a -
algebra.
-algebra/Field: Let X be non-empty set (space). A collection of subsets of X is -
algebra/field on X if and only if it satisfies the following properties:

1. X,
2. For all A , then AC
3. If An for n , then ⋃

Thus, a -algebra is a class of subsets of X that contains X and is closed under

complementation and countable unions.

Example 1: Let us consider X= {a, b, c} and the collection of subset is

*X * +* + +. Is -algebra?

9
Solution: By definition of -algebra we have

(i) X,
(ii) If X , then XC
, then
{a} , then * + * +
{b, c} , then * + * +
(iii) If the X, , {a}, and {b, c} is countable collection of sets in , then their union,
X * + * + X

Thus we can say that is a -algebra.

Home work:

1. If X= {a, b, c, d}, then the collection of subsets, *X * +* ++ and

* * ++. Are and -algebra? Justify your answer.
2. If X= {a, b, c, d} then show that the power set of X is -algebra.
3. If X = {a, b, c, d}, the collection of subset is defined as = {X, , {a, b}, {c, d}}. Is
-algebra?

Difference between algebra and -algebra:

-algebras are a subset of algebras in the sense that all σ-algebras are algebras, but not vice
versa. Algebras only require that they be closed under pairwise unions while σ-algebras must
be closed under countable infinite unions.

Largest -algebra: Let be the collection of the subset of a sample space/measurable space
X. If contains all possible subsets of X that is if is the power set of X, then is called
the largest -algebra or -field.

Example: Let us consider, X= {a, b}, then the power set of X is P (X) = {X, , {a}, {b}}, a
largest -algebra.

Smallest -algebra: Let be the collection of the subset of a sample space/measurable

space X. If contains only X and , then is called the smallest -algebra.

Example: The set = {X, + is the smallest -algebra.

10
Uses of -algebra in statistics:

1. In statistics, σ-algebras are needed for the formal mathematical definition of

a sufficient statistic. Particularly when the statistic is a function or a random process
and the notion of conditional density is not applicable.
2. In probability theory, where it is interpreted as the collection of events which can be
assigned probabilities. Also, in probability, σ-algebras are pivotal in the definition
of conditional expectation.

Borel -algebra: Let (X, ) be a topological space. We call the Borel -algebra (X) the
smallest -algebra of X containing .

Borel Sets: A Borel set is any set in a topological space that can be formed from open
sets (or, equivalently, from closed sets) through the operations of countable union,
countable intersection, and relative complement.

Measurable set: If A is -algebra of X, then X is called measurable space and the sets in A
are called the measurable sets in X.

Example: Let us consider X= {a, b, c}, then ={X, , {a}, {b}, {a, b}}. Thus is called
the measurable set.

Measurable space: A pair ( , ) is a measurable space if is a set and is a nonempty σ-

algebra of subsets of X. A measurable space allows us to define a function that assigns real
numbered values to the abstract elements of .

Measure: Let ( , ) be a measurable space. A set function μ defined on is called a

measure if and only if it has the following properties.

1. 0 ≤ μ (A) ≤ ∞for any A .

2. (Φ) = 0.
3. (σ-additivity): For any sequence of pairwise disjoint sets {An} such that
, we have μ ( An) = ∑ ( ).

Measure space: A measure space is a triple ( , ) where, is a set, is a -algebra of

the subsets of and is a measure on .

11
Theorem: Let ( ) be an arbitrary family of σ-algebra over S. Then show that their
intersection i.e is also a σ-algebra over S.

Solution: Since Fi; i=1, 2, …, n are σ-algebra over S, by the definition of σ-algebra over S.

S,∅ ; S, ∅ ; S, ∅ , ….., S, ∅

Hence, S, ∅ ⋂

Suppose A ⋂ , now we have to show that ⋂

Since A ⋂ , thus , , , …,

And since F1, F2, ……, Fn are σ-algebra, then we have

, ,….,

Thus, ⋂ .

Let A1, A2, ….., An ⋂ , we want to show that ⋃ ⋂

Since A1, A2, ….., An ⋂ , thus A1, A2, ….., An , A1, A2, ….., An , …., A1, A2,
….., An .

Therefore, A1 A2 ….. An , A1 A2 ….. An , …., A1 A2 ….. An .

Thus, ⋃ ⋂

All conditions of σ-algebra are satisfied by the intersection of σ-algebra.

Theorem: Show that the union of arbitrary number of σ-algebras may not be a σ-algebra.

Solution: Let be a measurable space and A1={A, Ac, ∅ } and A2={B, Bc, ∅ } be two
σ-algebra of .

It is enough to show that A1 A2 is not a σ-algebra of .

Since , ; thus and

If A1 A2 be a σ-algebra then A B must belongs to A1 A2, but A B A1 A2

i.e. A1 A2= {A, Ac, B, Bc, ∅}

Therefore, A1 A2 is not σ-algebra. Hence proved.

Home work:

12
 1. Let A1, A2 and is a σ-algebra. Do you think A1 A2 ? If so, display the
proof in favour of your answer.
2. Let A1, A2 and is a σ-algebra. Does ( ) ?

Elementary properties of measure

A positive measure function is a function , defined on a σ-algebra m which is range ( -

and which is countable additive. That means if * + is a disjoint countable collection of
members m then, μ (⋃ )=∑ ( ).

Theorem: Let (X, ; ) be a measure space. Show that for any A, B 2 , then show
that ( ) ( ) ( ) ( ).

Solution: Try yourself

Theorem: Let μ be a positive measure on a σ-algebra m, then show that

(i) (Φ) = 0
(ii) If ( ) = ( ) ( ) ( ), where are
pairwise disjoint members of σ-algebra in m.
(iii) If ( ) ( ) if
(iv) ( ) ( ) . If ⋃ , and

Solution:
(i) Since m is a σ-algebra and if , then by the property of positive measure

( )

Let

Then by the definition

μ (⋃ )=∑ ( )

( )= ( ) ( ) ( )

( )= ( ) ( ) ( )
13
 ( )= ( ) ( ) here, h is the total number of μ exceed of

( )

( ) ;h Hence proved.

(ii) We know that

μ (⋃ )=∑ ( )

( )= ( ) ( ) ( ) ( )

We put ∅, thus we have,

( ∅ ∅ )= ( ) ( ) ( ) (∅) (∅)

( )= ( ) ( ) ( )

Hence proved.

(iii) Since thus we can write

( ) and ( ) ∅
Now, ( ) ( ( ))
( ) ( )) ( )
Therefore, ( ) ( ).
Hence proved.

(iv) We put B1=A1 and

Then ∅
Since

If n=2, then we have,

If n=3, then we have,
Similarly, we can write,
Therefore,

14
 ( ) ( )
( ) ( ) ( ) ( )
( ) ( )

Again, Since ⋃ and

Now, ( ) ∑ ( ) ∑ ( ) ( )
( ) ( )
Hence proved.
Lebesgue measure: Consider the Real line R and the class of all bounded intervals of the
form (a, b] (a<b). On the elements of , we define a set function such that
(a) ( -
(b) (∅)
(c) (∑ ) ∑ ( )

Where , are disjoint sets of . Thus is a measure defined on . Since is closed

under finite intersection. can be extended uniquely to all Borel sets. This extended measure
on (R, ) is called the Lebesgue measure.

Outer measure: Let X be a set. A set function defined on the -algebra P (X) of all
subsets of X is called an outer measure on X if and only if it satisfies the following
conditions:

(a) ( ) , - (X)
(b) (∅)
(c) (X) ( ) ( )
(d) Countable sub-additivity: ( ) (X), (⋃ ) ∑ ( )

15
 CHAPTER THREE: RANDOM VARIABLE AND FUNCTION

Point function: Let us consider a mapping from to R. If the argument of function X are
points of space . Then this function is called point function. That is point function any
function whose values are point.

Example: Suppose a coin is tossed 3 times. The possible outcomes may be denoted by

* + * +
We may be interested only in the number of times head turns up. Thus the outcome
, we will associate a number X( ) , representing the number of head in . With the
outcome , we will associate X( ) , and so on.

Set function: Let us consider a mapping from to R. If the arguments of a function of X are
sets of a certain class, then this function is called set function. If is a class of sets, and with
each , we associate a value ( ), say, then is a set function. may represent such,
entities as weight, length, measure etc., associated with the sets of .

For example with an interval ( ) we may associate its length ; with the disjoint union
(( ) ( )) ( ) ( ) and so on.

Image, range, and domain of a function: A function X on a space to a space ʹ assigns

to each point a unique point in ʹ denoted by X ( ). X ( ) is the image of the
argument under X. It is also called the value of X at . X is also a mapping from to ʹ,
denoted by X . X( ) ʹ ʹ X

X also establishes correspondence relation between points of with points of ʹ. is called

the domain of X and ʹ is called the range. The set ʹʹ ,X( ) - which is a subset of
ʹ, is called the strict range of X. If ʹ ʹʹ, then we say that X is a mapping from

16
 ʹ, otherwise X is said to be a mapping from ʹ. The symbols
( ) X ( ) etc., will be used to denote functions, even though they really denote the values
of functions.

Inverse function: Let ʹ ʹ The set of all points of for which X( ) ʹ is called
inverse image of Bʹ under X, denoted by X ( ʹ):

X ( ʹ) , X( ) ʹ-

Thus with every point function X, we associate a set function X , whose domain is a class
ʹ of subsets of ʹ and whose range of class , (say) of subsets .X is called inverse
function (or mapping) of X. That is

X( ) ,X( ) -
X ( ʹ) ,X ( ʹ) ʹ ʹ-
Example: Consider the function X (w) =w2 from R to R. Let B be the set of points contained
in the interval (1, 2). Then X ( ) ( √ ) ( √ ). If the range and the domain of X

is the same and is ( ), then X ( ) ( √ ).

Lemma: Inverse mapping preserve all set relations.

Proof:

(i) Let ʹ. Then we have

X ( ) , X( ) - , X( ) - X ( )
X ( ) X ( )
(ii) Let ʹ, then we have
X ( )
X( )
X ( )
X ( )
X ( ) X ( )

X ( ) X ( )
(iii) X ( )
17
 X( )
X( )
X ( )

(X ( ))

Thus, X ( ) (X ( ))
Evidently, X ( ʹ) , X( ) ʹ- X ( )
If
X ( ) X ( ) X ( ) . Hence X ( ) and X ( ) are disjoint.
Corollary: If is a class of subsets of and is a -field, then show that the class of of all
sets whose inverse images belong to is also a -field.

Proof:

By definition of , we have
X ( )X ( )

⋂X ( )

X (⋂ )

Hence is closed under countable intersection.

Again, If
X ( )

(X ( )) .As is a -field
X ( ) , (X ( )) X ( )-

Thus, is also closed under complement. Hence a -field.

Corollary:
(i) If C is a field ( -field) of subsets of ʹ X ( ) is a field ( -field) of subsets of .

18
 (ii) Inverse image of the minimal -field over any class C is the minimal -field over
X ( )
*X ( )+ X * ( )+
Proof:
(i) Let C be a -field on ʹ and be the class of inverse images of the elements of C.
Then, C is a -field.
X ( )X ( )

X (⋂ )

⋂X ( )

Thus, is closed under countable intersections.

Again, if X ( ) .
.
. Since C is a -field.
( )

( ( )) Using ( )=( ( ))
Thus is closed under complementation and hence it is a -field.
Therefore, the inverse of -field is also a -field and in particular, inverse of a field is a field.

Indicator functions: Consider a real-valued function or I (A) defined on as follows:

( ) 2

Then is called the indicator function of A. The strict range of is

( ) * ( ) + * +
Measurable function and Borel measurable function: A function X is said to be
measurable functions if it satisfy the following two conditions:
(i) X ( ) ( ʹ ); Where and are -field on and ʹ with respectively.
(ii) X ( ) , for all Borel set . Equivalently, X ( ) .

19
If real line and in a Borel field on in the above conditions. Then X
is called -measurable function or a function measurable with respect to .
If real line and , the Borel field on in the above conditions. Then X is
said to be the Borel function.

Theorem: Let be an indicator function defined by

( ) ( );
Then is a measurable function if and only if .
Proof: Let be a measurable function and also let , then

( ) {

Thus ( ) * + ( ) which is -field.

Thus we see that ( ) .
Conversely, let . Since is a -field on , then we have
and also
* +
is a -measurable function.
Example: Let us consider a function
X:( ) ( ʹ ); ( );
* +; ʹ * + And
* + Shown as follows:

X( ) {

Show that X is a measurable function.

Proof: Let . Thus

20
 X ( )

{
Hence
X ( ) * +
Which is the minimal -field containing the partition * +. Thus X is a measurable
function.
Simple function: A finite linear combination of indicators of sets (not necessary disjoints) is
called a simple function. If is finite, all real valued functions will be simple.
Lemma: Any simple function can be written as

X ∑X

Where, X are distinct numerical constants and * + is a partition of .

Proof: Let X be a linear function of only, say
X
Then

X( )

{
Hence X( ) * +. Moreover,
* + forms a partition of . Thus the lemma is proved for the linear combination of
two indicators.
Suppose in general, X is a linear combination of m indicators

X ∑

Then X( ) . If .
Where, ( ) is permutation of (1, 2, …, m) and k=0, 1, ….,m. Thus X can have at
most 2m values which are distinct. The set being disjoint * + form a partition of . The
lemma follows by taking n, the number of distinct values of X.
21
 -field induced by X: Let X be a real valued function on and be the Borel field. The class
of sets X ( ) *X ( ) + is a field. It is called the -field induced by X. It is also
denoted by (X) X will be -measurable if and only if it induces a -field which is a -field
of .
Function of a function: If X is a function from to ʹ and Xʹ is a function from ʹ ʹʹ,
then X(Xʹ( )) is a function from to ʹʹ, denoted by XʹX or Xʹ(X). It is said to be a function
of function or a composition of two functions X and Xʹ. It inverse is a function on the subsets
of ʹʹ ʹ such that for any B ʹʹ

(XʹX) ( ) [ Xʹ(X( )) ] [ X( ) Xʹ ( )] X .Xʹ ( )/

(XʹX) X Xʹ .

Random variable: Let ( )be a probability space. Then a function: is called a

random variable, if the event
X (( -) * X( ) +
For each a in R, i.e., a random variable is a real valued -measurable function on a probability
space ( ).
Economical definition of random variable: Let be the -field of events associated with a
certain fixed experiment. Any real value measurable function defined on is called a
random variable. Thus X is a random variable if and only if X ( ), the -field induced by X
is contained in or equivalently, X ( ) is an event for every Borel set B in the range space
of X.
Example: A simple function X ∑ X induced by -field containing the partition
* +. If the partition is contained i.e., if each one of the Ai‟s is an event, then
X is a random variable called a simple random variable. Since singleton * + ( ) belongs to
. That is
(X ) , X( ) - X (* +)
Again, Borel function ( ) of a random variable X is also a random variable and induces a -
field which is induced a sub- -field of that which is induced by X. For a given X, define two
non-negative functions:
X( ) X( )
X ( ) {
X( )
And

22
 X( ) X( )
X ( ) {
X( )
Called respectively, the positive and negative part of X. Then X and X are Borel function of
X and will be random variable if X is a random variable. We may note that X and X are non-
negative functions such that for all w,
X( ) X ( ) X ( )
|X( )| X ( ) X ( )

Lemma: X is a random variable if and only if X ( ) , where is any class of subsets of

R, which generates by .
Proof: We have to prove that
X ( ) X ( )
Since ,X ( ) ,X ( ) To prove the converse, Since is a -field and
X ( ) (X ( ))
X ( ( ))
X ( )
Vector random variable: With any associate ( ) (X( ) ( )), a point in two-
dimensional Euclidean R2, Z defines a function from . In , consider the class of all
rectangles bounded by the lines x=a, x=b, y=c, and y=d, a<b, c<d are the arbitrary. The
minimal -field containing is Borel field in . It is also the minimal -field containing
all finite rectangles of the form ,( ) - Z will be called a two-
dimensional random variable if ( ) ( ) is the -field induced by Z. In analogy
with the economical definition, Z will be random variable if and only if ( ) is an event for
any rectangle in .
Example: Let , -. Then
( )
( )
( )
( )
{( )
Now,

23
 ( )
( ) ( )
( )
{ ( )
Hence ( ) * + ( ) ( )
Thus Z will be a random variable if A and B are events.

SECTION -B

CHAPTER FOUR: INTEGRAL OF MEASURABLE FUNCTION

24
 CHAPTER FIVE: PROBABILITY MEASURE

Axioms of probability:
1. ( ) for every event .
2. ( )
3. Let A1, A2, …, An be a sequence of mutually exclusive events in , then

( ) ∑ ( )

4. If A1 and A2 are two mutually exclusive events in , then

( )= ( ) ( )

Some simple properties of probability:

1. The probability of an impossible event is zero. That is ( )

2. The range of the probability of the event A lies between 0 and 1. That is ( )
.
3. If A1, A2, …, An be a sequence of mutually exclusive events in , then

(⋃ ) ∑ ( )

4. If A1 and A2 are the subset of such that , then ( ) ( )

Probability measure: Let ( , ) be a measurable space. A set function P defined on is

called a probability measure if and only if it has the following conditions.

1. 0 ≤ P (A) ≤ 1 for any A .

2. ( )
3. If , i=1, 2,.., and Ai and Aj are disjoint, that is ∅
P (⋃ An ) = ∑ ( ).

The triple ( , ) is called probability space.

Theorem: Show that (⋃ )=∑ ( ), where are pairwise disjoint.

25
Solution: Try yourself

Discrete probability space: A discrete probability space is a triplet ( , ) such that

(a) The sample space is finite or countable. That is * +

(b) The σ-algebra is the set of all subsets of .
(c) The probability measure assigns a number in the set [0, 1] to every subset of It is
defined in terms of the probability space P (w) of the elementary outcomes and
satisfies:

( ) ∑ ( )

, and

∑ ( )

Example: Consider a single toss of a coin. If we believe that heads (H) and tails (T) are
equally likely, the following is an appropriate model. We set * +. Where,
and let P1=P2= . Here, *∅ * + * + * ++ and (∅) ( )

( ) , P (H, T) =

Properties of probability measure:

(a) Finite additivity: If the events are pairwise disjoint, then

(⋃ )=∑ ( )
(b) For every event A, we have ( )=1- ( )
(c) If A and B are the subset of such that , then ( ) ( )
(d) Union bound: If A1, A2, …, An be a sequence of mutually exclusive events in , then

(⋃ ) ∑ ( )

Theorem: Prove the following properties of the probability:-

(a) If A1, A2, …, An are disjoint, then show that (⋃ ) ∑ ( )

26
 (b) Let A be an event, then show that ( )=1- ( )
(c) If the events A and B satisfy then ( ) ( ), where P is probability
measure.

Solution: Try yourself.

Theorem: Under usual notations, prove that if ( ) ( ) and if

( ) ( ).

Solution: Let * + be a monotone increasing sequence, then we have .

Now, ⋃

( ) ( ) ( )

∑( )

( ) ( ) (∑( ))

Taking limit as

( ) ( ) (∑( ))

( ) ( )

( ) ( ). Hence proved.

2nd Part:

Let * + be a monotone decreasing sequence, then we have , then we have

( ) ( )
27
 ( ) ( )

( ) ( )

( ) ( ). Hence proved.

Legbesgue-steieltje’s measure: Let g be any monotone non-decreasing point function on R,

which is finite or having bounded variation. Without the loss of generality, it may be taken to
be continuous on the right (left), by taking the value of the function at a jump point to be the
limit of its values from the right (left). Thus ( ) ( ) ( ) .
On the sets of defined a measure as follows:

(a) ( - ( ) ( ) ( -
(b) ( ) ( ) ( )
(c) (∑ ) ∑ ( )

Where * + is a class of disjoint set of . Since is non-negative and continuous on , it

can be extended to all Borel sets and hence it is measure on (R, ). This is called Legbesgue-
Steieltje‟s (LS) measure induced by the function g.

Signed measure: Let (X, ) be a measurable space. A signed measure on (X, ) is a

function µ: → Rn such that

(a) µ takes on at most one of the values

(b) µ(∅) = 0
(c) If * + is a sequence of pairwise disjoint sets, then

(⋃ ) ∑ ( )

Example: Assume that (X, A) is a measurable space and that µ1, µ2 are measures on (X, ).
If at least one of the µi‟s is finite, then µ = µ1 − µ2 is a signed measure on (X, ).

28
Zero-one law: In probability theory, a zero–one law is a result that states that an event must
have probability 0 or 1 and no intermediate value. Sometimes, the statement is that the limit
of certain probabilities must be 0 or 1. It may refer to:

 Borel–Cantelli lemma.
 Kolmogorov's zero–one law for the tail σ-algebra.
 Blumenthal's zero–one law for Markov processes.
 Engelbert–Schmidt zero–one law for continuous, nondecreasing additive functionals
of Brownian motion.
 Hewitt–Savage zero–one law for exchangeable sequences.
 Lévy's zero–one law, related to martingale convergence.
 Topological zero–one law related to meager sets.

Theorem: State and prove the Borel–Cantelli lemma.

Statement: Let * + be a sequence of events over a probability space ( , ). Then

∑ ( ) ( )

If * + are pairwise independent, then also

∑ ( ) ( )

Proof: Let * + be a sequence of events over a probability space ( , ) and suppose

the sum of probabilities of the An is finite that is

∑ ( )

Let us define
A= ⋂ ⋃
This implies

And thus

29
 ( ) (⋃ ) ∑ ( )

Since ∑ ( ) converges, then ∑ ( ) converges to zero as . Thus P (A) =0.

Hence proved.

2nd part: Let us define A= and furthermore

And eventually

Since the are pairwise independent they are pairwise uncorrelated as well. Hence

( ) ∑ ( ) ∑0 ( ) { ( )} 1

( ) ∑ ( ) ∑{ ( )}

Since , then we have

( ) ∑ ( ) ∑{ ( )}

( ) (∑ ) ∑{ ( )} ( ) ∑{ ( )} ( )

Now by the assumption ∑ ( )

( ) ( )

On the other hand if and only if for in finitely many n which is the case, if and
only if ( ) = . Thus the assertion is
( ) ( ) .
Hence proved.

30
 CHAPTER SIX: DISTRIBUTION FUNCTIONS AND EXPECTATION

Distribution function: A real-valued function F defined on ( ) that is non-decreasing

and satisfies
( ) ( )
is called a distribution function (DF).
Alternative definition of DF: Let X be a random variable defined on ( ). Define a
point function ( ) on R namely,
( ) ( - ( X( ) )
The function F is called the distribution function of the random variable x.
Example: Let X=I (A) with P (A) =p, then the DF is

( ) {

Since X takes the values 0 and 1 only, with the probabilities (1-p) and p respectively. F(x) has
jumps of magnitude (1-p) and p at 0 and 1 respectively.
Conversely, if F(x) has jumps of magnitude p and 1-p at a and b respectively, it is the DF of a
random variable (RV) taking the values a and b with probabilities p and 1-p respectively. If
F(x) has a jump of magnitude unity at x=c, it is the DF of a degenerate RV X, which is a
constant equal to c.
Properties of DF:
1. ( )
2. ( )
3. F (x) is an non-decreasing function. That i then the DF is
( - , - ( ) ( )
4. F(x) is right continuous, i.e.,
( ) ( )

Probability distribution of a random variable

Definition/Theorem: The random variable X defined on the probability space ( )
induces a probability space ( ) by means of the correspondence
( ) ,X ( )- , X( ) -

31
We write Q= X X the probability distribution of X.
Proof: Clearly,
( ) and
Also ( ) ,X - ( )
Let . Since the inverse image of a disjoint union
of Borel sets is the disjoint union of their inverse images. We have

(∑ ) [X (∑ )]

[∑ X ( )]

∑ X ( )

∑ ( )

It follows that ( ) is probability space and the proof is complete.

Theorem: The function F defined as
( ) ( - ( X( ) )
is indeed a distribution function (DF).
Proof: Let . Then , - , -.
Now,
( ) (X )
(X )
( )
Since F is non-decreasing, it is sufficient to show that for any sequence of numbers X X.
That is X X X X., i.e., (X ) (X).
Let * X( ) (X X -+. Then . Also ⋂ .
Since none of the interval (X X - contains X. So that (X ) (X).
That F is right continuous.
Finally, let *X + be a sequence of non-decreasing . Then we have
*X X + *X X +
And *X X + ⋂ *X X +

32
 ( ) (X X )

0 (X X )1

, -

Similarly, ( ) (X X )
Hence, the theorem is proved.
Theorem: Every DF is the DF of a random variable on the same probability space.
Proof: Let F be the distribution function. It follows that there exists a unique probability on R
that satisfies
( - ( )
Let ( ) be the probability space on R which we define
X( )
Then
( X( ) ) ( - ( )
And F is the distribution function of the random variable.
Hence the theorem is proved.
Example: Let X be defined on ( ) by
( )
Find F(x).
Solution: We have
(X )
( - *X ( - +
( ) {

Example: Let * + X
X( ) X( )
( )
Solution: If p assigns equal mass to * + * +, then we have

( ) ( )

And

33
 ( ) ( - {

Example: Let *( ) + be the set of all subsets of Let * +

for all pairs (i, j); . Define X( ) Find F(x) and

draw it.
Solution: Try yourself.

Probability mass function: Let * + be a collection of non-negative real numbers such that
∑ . Then * + is called the probability mass function of some random variable x.
Probability density function: Let X be a random variable on ( ) with distribution
function F (.). Then X is said to continuous type if F is absolutely continuous that is if there
exists a non-negative function F(x) such that for every real number x, we have

( ) ∫ ()

The function f(x) is called probability density function of the random variable x.
Note that: ( )

( ) ( ) ∫ ( )

Let a and b be two any real numbers .Then we have

, - ( ) ( ) ∫ ()

Theorem: Let X be a random variable of the continuous type with pdf “f” then for every
Borel set

( ) ∫ ()

If F is absolutely continuous and f is continuous at x, then we have

( )
( ) ʹ( )

Theorem: Every non-negative real function f that is integrable on R and satisfies

34
 ∫ ()

is the pdf of some continuous random variable x.

Proof: It is sufficient to show that there corresponds a distribution function F to f.
We define

( ) ∫ ()

Then ( ) ( ) , then we get

( ) ∫ () ∫ ()

( ) ( )
( ) ( )
Finally, F is (absolutely) continuous and hence continuous from right.
Hence the theorem is proved.
Example: Let X have the triangular pdf

( ) 2

Find (i) F(x); (ii) ( ) ( ) ( )

Solution: Try yourself.

Function of random variable:
Definition/Theorem: Let X be a random variable defined on ( ). Also let g be a Borel
measurable function on R. Then g(x) is also a random variable.
Proof: For
* ( ) + * ( -+
Since g is a Borel measurable, then ( - is also a Borel set. It follows that * ( )
+ and the proof is complete.

35
Example: Let X be a random variable with DF F. Then |X|, X ( ),
X ( ) |X| ( ) .
Solution: Define

X {

And

X {

Then X and X are also random variables. We have

(| | ) ( ) ( ) ( ) ( )
Thus | | is a random variable.

( ) ( ) ( )

( )
( ) {
( )

Now, (X ) { ( )
( ) ( )

Similarly, (X ) {
( )
Theorem: Let X be a random variable of continuous type with pdf f (.) and y=g(x) be
differentiable for x and either gʹ(x)<0 for all x or gʹ(x)>0 for all x. Then y=g(x) is also a
random variable of continuous type with pdf is given by:

, ( )- | ( )|
( ) {

Where, , ( ) ( )- , ( ) ( )-
Proof: If g is differentiable for all x and gʹ(x)>0, then g is continuous and strictly increasing
limits exists and the inverse function x= ( ) exists is also strictly increasing and
differentiable. The distribution function of y is given by
( ) ,X ( )-

The pdf of g is then obtained and differentiations

36
 ( ) ( )

( ( ))

, ( )- ( ( ))

Similarly, if gʹ(x)<0, then g is strictly decreasing and we have

( ) ,X ( )-
,X ( )-
So that

( ) , ( )- ( ( ))

Since g and g-1 are both strictly decreasing ( ( )) is negative and the result follows

, ( )- | ( )|
( ) {

Expectation: Let X be a simple random variable

X ∑ ( ) ∑

Its expectation is denoted by EX and is defined as

X ∑ ∑ ,X -

It is also called the integral of X with respect to P-measure and is denoted by

X ∫X

Properties of Expectation:
1. Linearity: If X and Y are simple random variables, then
(X ) X
2. Scale preserving: If c is a constant, then ( X) X
3. Non-negativity: If X X

37
Theorem: If X and Y are simple random variables, then (X ) X .
Theorem: If c is a constant, then ( X) X.
Theorem: If X X .
Theorem: If X X .

38
 Assignment

1. Define lebesgue integral. Write down the drawback of Riemann integral in terms of
Lebesgue integral.
2. Write down the properties of Lebesgue integral.
3. If f is bounded function defined on [a, b]and f is Riemann (R)and Lebesgue (L)
integrable on [a, b], then show that

∫ ( ) ∫ ( )

4. Let f be a bounded function defined on [a, b] and let f be Lebesgue (L) integrable over
[a, b] and [c, b], then show that

∫ ( ) ∫ ( ) ∫ ( )

5. Define Lp-spaces. Let ( ). Then prove that for any

∫( ) ∫ ∫

39