h a n g e Pro h a n g e Pro

XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Introduction
to Probability
Session 1

1
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
2
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Aims and Focus

 Our Focus is on basic ideas and concepts
 Understanding of formulas and applying
them in real problems of your research
 We need to choose appropriate formula
and make judgments
 Many solutions exist for a problem but
shorten one needs deeper understanding
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
3
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Importance Reasons
 Mostthings in the life are uncertain
 Way of dealing with them is developing
models based on the probability theory
 Examples:
 Weather forecasting
 Service centers
 Stock market
 Basic Ideas are same in all majors
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
4
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Agenda for
Basics of probability
 Probability is a framework for reasoning
about uncertainty
 Probabilistic models:
 Sample space
 Probability law
 Axiomsof probability
 Examples
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
5
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Sample space 
 Set of possible outcomes of a random
experiment
 Set must be:
 Collectively exhaustive
 Mutually Exclusive
 Tradeoff between level of details that
must be considered in the model is
experience
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
6
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Sample Space: Discrete

Example
• Two rolls of a tetrahedral die as a random Stage 2
experiment  twice repetition of an Stage 1
(1,1) outcome
experiment  two different experiments 1 (1,2)
• Sample space vs. sequential description (1,3)
(1,4)

2
X = First Roll

1 2 3 4
1
Y = Second Roll

2 (3,4)
3 (4,3)
4
4 (3,4) (4,3)
(4,4)
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
7
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Sample space: Continuous

example
 Infinitesample space
  = { (x, y), 0  x, y  1 }

0 1 X
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
8
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Probability Axioms
 Event: A subset of sample space 
 Probability is assigned to events (sets)
 Probability :belief that how much A
something is likely to occur
 Axioms of probability
 Non negativity: P(A) ≥ 0
 Normalization: P() = 1
 Additivity: if AB = , then P(AB) = P(A)+P(B)
(disjoint events)
 These axioms are sufficient
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
9
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Some conclusion: argument

(2) (3)
 1= P() = P( A  Ac ) = P(A) + P(Ac) →
(1)  A
P(A) = 1- P(A )  1
c
(3) (3)
C B
 P(ABC) = P(AB)+P(C) =
P(A)+P(B)+P(C)
 Let A1, A2, …, An are finite number of sets
 P(A1A2…An) = P(A1)+P(A2)+…+P(An)
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
10
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Conclusion
 P({s1,s2,…,sn})=P({s1})+P({s2})+…+P({sn}) =

 s1
P(s1)+P(s2)+…+P(sn)
s2 sn

 Only for finite countable sets.

 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
11
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Probability low: Example with

finite sample space
 Let dies are fair X = First Roll

 P((X,Y) is (1,1) or (1,2))= 1 2 3 4

2/16 1

Y = Second Roll
 P({x=1})= 2
4/16 3
 P(X+Y is odd)=
4
8/16
 P(min(X,Y)=2) =
5/16
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
12
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Discrete Uniform Low

 Let assume that all outcomes are equally
likely
number of elements of A (𝑚) 1
𝑃 𝐴 = = 𝑚
total number of sample points(𝑛) 𝑛
 Computing probabilities = counting
 Defines fair coins, fair dies, etc
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
13
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Continuous Uniform Low

 Dart Example: Two “random” Y
numbers in [0,1] 1
 Uniform Low: Probability = area
 P((X,Y)=(0.5,0.25))=
1/2
Zero
 P(X+Y1/2)=
1/8
1 X
1/2
X+Y=1/2
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
14
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Probability Low: Ex. W/countably

infinite sample space
 Probability of seeing first head in nth coin flip
 Sample space: {1,2,3,…}
1/2
 P(n)=2-n , n=1,2,3…
 Find P( outcome is even)
1/4

Countable additivity axiom: 1/8

if A1, A2, …, An are disjoint 1/16
…
events, then
1 2 3 4
P(A1A2…An)=P(A1)+P(A2)
+…P(An)
1 1 1 1
𝑃 2,4, … =𝑃 2 +𝑃 4 +⋯= + +⋯= 4 =
4 16 1−1 4
3
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
15
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Counter Example
Y
 P(A)= area(A) 1
A

□= { 𝑥, 𝑦 }
𝑥,𝑦

1 X
1=𝑃 □ =𝑃
𝑥,𝑦
𝑥, 𝑦 =
𝑥,𝑦
𝑃 𝑥, 𝑦 = 0=0

∞
• Mistake is in applying Additivity axiom
• Countable infinite set  uncountable sets { 𝑥, 𝑦 } 𝐴𝑖
𝑥,𝑦 𝑖=1
 h a n g e Pro h a n g e Pro
XC d XC d
F- F-

uc

uc
PD

PD
!

!
W

W
t

t
O

O
N

N
y

y
bu

bu
16
to

to
ww

ww
om

om
k

k
lic

lic
C

C
.c

.c
w

w
tr re tr re
.

.
ac ac
k e r- s o ft w a k e r- s o ft w a

Some results
 Zero probability does not mean impossible
 P( (x,y)(0,0) ) = 1
 Probability of one means essential
certainty but still other outcomes is
possible