x =

4
Alvaro Domenech
Ne

Density function X no
Distribution function =
events pen unit of time
dF(t) X e-
**
+ 20
F( 1 +
= PIT = +) 1 e
- X t
. = .

for T = time until 1 event occurs

dt
= -

Des
All

Properties related with Poisson process

P(X
Properties :

A 1 1)
=1 A=
Des
.
= =
p
1
Properties times something has 2 P(x
.
1 E[ + ] = X
X =
happened .
=
0) = 1 -

p
no defective items Possible outcomes 3 E(x) time 2 Van [ +] 1x
.
1 something
=
X = Exi -
independent variables n =
.
=
p btw ·
.

.
3
2 E(X) E([xi) EE(xi) 4 Van (x) p (1 p) * Has no memory
probability Exponential Random
n
p p
= = = - .
= .
-
.
=
Agro
3 Van (x) =
Van([xi) =
[Van(xi) = n p q
.

scaling factor
.

mean
M
.
= I

XvU(a b) Variable
.
4 If X1 and Xc are independent :

Distribution function
, large
noof random
processes O = sta deviation =
translating faton
x1 + x2vB)n1 + nz p) XvB(n p)
De
all
Properties
,

#
, :

-Xb
-
A
if a
E(x)
f(x) Median
Lob
b .
1 =

m = = Mode
In any other case ·
2 Van(x) = 0
.

Bernoulli's
mass function
k Experiment Uniform Random
Normal Random
3
. Linear transformation of
(2) (1 p)" normal variable random
a -

ph
variable
p(k) P(X k) Variable
.
-

= = = .

E
3

for k = 0
,
1
,
2
, ...
n
Binomial distribution
Properties Al Variable · basis for statistical inference
. +
1 E(x) =
a
(b-all
.

-
(x -
y)
1 20
2 Van (x)
1(x
=
.
· = -
e -
02-

YuG(p) =

P(y k) = = (1 -

p) a. p big amount of data > 30

-

Discrete Continuous Central Limit &

distribution
Geometric Random Models Probability Functions ·

Models
Da
normal

E
Theorem
Properties :

Variable used to choose sample size

B ·

=
before investigation
1 .
E(Y) =
- 2 .
Van (y) F

a
mean distribution more similar
YIN Mi I &

.
3 Has memory
,
i
to
no of a
sample normal

aff
②

Hypergeometric
N n in
population Random
=

sta
.
error error

Poisson
=

n =
sample size as a Binomial

To
K no
of events in the event
Variable
=

approximation
Desp
k = no of events in the sample B Binomial Normal higher · better normal estimation

C
N -
K
Poisson to Normal lower sta
Cn-k
approximation
.
error
P(X = k) =

CN
Interval
approximation
D
of time
·
Discrete
· Continuous
Poisson Random Th B
Rare events
repeated
lots
& Variable if pc0 1 and n p > 5 X = n .

p
of
.

times
.

When I
n > 30 ·
p(X x) = = p(X = x + 0 5) .

B ·
Correction
Poisson
n .

p .
/1 - p) > 5

X-P(X)a
arises as
·
n tends to N
factor ·
p(x = X) = p(x -
0 .
3 - X)
limit of binomial
e

D Y*N n
p, n
p (1 p)
. -
. -

X>5 mean
G
X = usual no elements
pen unit of time
e Apply correction

X no of observed
Properties :
probability function factor
= events in each unit of time &
N

1 No
.

memory
e 10 5 X-p(X)wNX , X
2 . E(y) = Van (y) = X- Stable mean p(k) = P(y =
k) = -
.

.
3 Y = X1 + Xz

Y vp(Xn Xz)
up(xi))
X, E
+

XzwP(Xz (