Probability (basic

Basic Definitions

·
Union of A and B , CAUB) :
the probability of A OR B
occurring

Intersection of A and B , (A & B) the probability of BOTH A and B

occurring
·
:

no . of outcomes in A
P(A)
sample space
· =

total no
. of outcomes
notation Multiplication Rule
·

-
=

·
P(A)20 for all events ALE P(AnB) = P(B) x P(A/B)

·
Pr(E) = 1

Law of total probability

·

Pr($) 0
, A representing empty P(A(B) x P(B) + P(a(B') xP(B)
· =
set
P(A)
.

·
Not'A' >
-
Pr(A)) = 1 -

P(a) (complement)

·
Pr( = VB) = Pr(A) + Pr(A1B) -> addition rule

Mutually exclusive :
nothing in common >
-
events do not add up to 1

if events A and B are

mutually exclusive
,

Pr(A1B) = 0 Pr(AUB) = P(A) + P(R)

,

Independent Probability Conditional Probability

keywords : GIVEN , If A and B
-
·

Pr(A (B) =
Pr(A)
Pr(B(A) Pr(B) P(AnB)
P(A(B)
·
=
=
-
·

Pr(A (B) = P(A)xP(B) P(B)

'A B' B
given =

Probability Table Tree

Diagram
Stage 1 Stage 2 Outcomes
H H H P( + ) xP(t)
B B'
H
A AlBAMB' P(A) P( + ) xP(T)
T HT
A'ARB Two Coins
A'RB' PCA')
P(B) P(B') 1 H
TH
P(T) xP( + )

T
P(T) x P(T)
T TT
Discrete Random Variables

Two conditions :

. Each value of p(x)

1
belongs to the interval [0 1)
,
>
-
0 = P = 1 for all >

. The sum of all the values of

2 p(x) must be 1 >
-
[p(r) = 1

Discrete : a countable number of possible values

>
-
probability of each value is between O and I

1
>
- SUM of all probabilities =

Expected Value (mean)

Notation : E(X) or M
>
-
sum the products of each value of X
and the probability that X takes that value
.

u =
E(X) = 2xxP(X = x)
2

[x
=
xp(x)
x

E(X) of a function of X :

E(g(x)) =

2g(x) xp(x)
E(X + Y) =
E(X) +
E(y)

E(X) of aX + b -
E(aX + b) = aE(X) + b

(for a
,
b constant)

Variance & Standard Deviation

Variance (ah) spread of probability distribution
:

about its mean

E((X u)2]
2
o =
var(X) = -

or var(x) = E(x2) -

[E(X)]2

standard deviation :

o = sd(X) =
vvar(X)

Var(aX + b) Var(x)
=
Linear function >
- = a

* In
general for many
,
random variables X,

Pr(m -
20 = X = m + 20) = 0 . 95
Bernoulli Sequence
· success or failure (two outcomes)
·
p
=
probability of success on a
single trial (failure = 1 -
p)
·
trials are independent

Binomial random variable : the number of successes in a Bernoulli sequence .

Binomial Distribution : Pr(X x) = =

(n)p" (l - p)" -
n !

where (2) =
x ! (n -
x) !

Notation : XvBi(n p) ,
>
-

n and pare the parameters of the distribution

where
,
n = no .
of trials Mean =
M= E(x) =
nxP

p
=
probability of success
-
7 Variance 02
= =
nxp(1 -

P)
1- P =
probability of other outcome Standard Deviation = o = vo
(failure)

Finding the Sample Size (n)

Inverse Binomial N Rules about Binomial .

·
(8) =
1 ·
(Y) = n

Pr(X21) > 0 95 .
·
(n) = 1 -
(n 1)
-
= n

1 -
Pr(X 0) =
>0 95 .

Pr(X 0) <= 0 05.

# · (0 52)" < 0 .
05
.

0 .
52" < 0 05 .

n
= 4 58 .

.. n = 5

Pr(X= 2) 70 .
95

1 -
Pr(X(2)70 95 .

Pr(X(2) < 0 05

%0 48' (0 32)"
.

40 (0 52)" +
05
# .
.
10 .

-
0 .
524+ 0 .
48n (0 .
52)" 10 . 05

n
= 7 . 7986

·
. n = S
-
Continuous Random Variables
f(x
one that can take any value . probability

M
-
* total area = 1
Probability density function
:

F(x) = P(a = x = b) =

f f(x)dx = 0 O a b
-x

P(a(x = b) =
p(a = x(b) = P(acx = b) =
P(a = x = b)

for x[a b] ,
Sf(x)di = 1

·
for xe[a c) ,
(f(x)dx = 1

Forxe[a b) P(xx) S f(x)dx

&
· =
,
,
P(xx) +
P(x)x) =
7
P(x() =

(P,f(x)dx

P D F must
A . .

satisfy 2
properties :

>
-
f(x) = O for all I

>
-
ff(x)dx =
1

2
Variance and Standard Deviation Var(x) =
E(X) -
E(X)" = o

·
68 % of data = 1S D .
of mean
. P(M -

0 < XM +
0) = 0 68.

·
95 % of data = 2S D of . .
mean P(M-20X = M + 20) = 0 .
95

· 99 7 % of data
. = 3 S D .
of mean . ↑) (M 30
-
= X =M + 30) = 0 . 997

(b) (P)
Mean M = E(X) =
S xx f(x)d IQR = 75th percentile - 25th percentile

· 25th percentive =
(f(x) =
0 25
.

E(g(()) =
(g(x)xf(x)dx
·
75th
percentile =
Sf(x) = 0 . 15
·
E(Xz) (x2xf(x)dx
=

·
E(aX + b) = af(x) + b

Median P(X [ M) = P(X ) M) = 0 . 5

0 5 .
=
Sw f(x)dx = P(X = M) =
P(X)M)
 Normal Distribution

1 -2 ) =)
f(x) =

oven
Ye for -
>x >

·
Notation : X-N(M , 02)

The 68-95-99 7 Rule .

Properties :
·
Extends to co in both directions

·
curve symmetrical about the mean .

Pr(axx(b) =

S f(x)dx

Area =
S % f(x) do
.
= 1

↑ Standard Normal dist : zN(0 1) ,

+u = 0
,
0 = 1 z =
Populations and
Samples
6 (seed)
Generating sample CAS 5-> 4 -
a :

5 - 4 -
1
5 - 4 - 2 randInt (lower bound, upper bound, no of results

no . of population wh attribute
Population proportion p =
population size

no .
in sample with attribute

sample proportion P =

sample size
(na+ 1)

small Population Hypergeometric

-
Distribution (discrete)

distribution : !
Hypergeometric (3) =
n
is the
D = # desired characteristic in x ! (n-1) !
pop
(a)x( 2)
.

=
d = # desired characteristic in sample no of ways to select
P(X d)
.

=
=

N pop size .
objects from n objects
(i)
= .
x

n =
sample size

Distribution
Sampling

↑
01
P( =
p) t 0

Mean and Standard Deviation from small population

E(p) =
[(pXP(p)) ·
SD(F) VE(P2) = -
E(p)2

LARGE population .
E(P) =
p(pop proportion)
.

,
sa(i) =
Large pop Small .
sample >
- Binomial distribution .

distribution
Large pop Large Sample Normal
.
>
-
Confidence Intervals for the Population Proportion
·
p(pop proportion)
. is UNKNOWN.

Point Estimates : sample proportion & is an estimate of P .

Interval Estimates - a confidence interval for p.

-
I

90 % C .
I 1 . 65 note : E values give the

level of confidence
.

997
95 % C I
.

C(upper) + (Clower)
= 2

CAS : menu >

- 6 >
- 6 >
- 5

Finding the Escore -

just solve func .