Lecture : April ?

"a
3 continuous RVS

1. uniform
2 .
exp (xp(X)
3 .
normal/gaussian N(M , 0

examples
discrete poisson RV- # =
# of people entering

from 9-10 AM

continuous exp RV- > I = how long until the

next person enters

continuous normal RV-Z = amount of error

when measuring something

suppose I denotes the number of events in a given continuous

interval . Then I follows an approximate Poisson RV with parameter

XO .
If

9 .
# of events occuring in non-overlapping intervals are independent

b
. prob of exactly one event occuring in a short interval I/M is

X
th
:

approx

of events in short interval

C .
prob exactly two or more occuring a

is zero
 Lecture : April ?"a
Deriving the pmf for Erpoisson (1)

# = # of events in the interval (0 , 1)

in the interval
on average , events occur

split into
several
the interval

probability of this event is I

R

event 1-X/n
no
happening

P(I = 1) =
(d)(2)(1-n-K for large o

limit n> a
-

im -
=
long proof

Xe
>
probability that a

=
-

#= 1)

Are
↑ (1)) =
P(poisson(x) = k) =
 Lecture : April ?"a

f(x) =
P(Bern(p) = 1) =

Si
n1
1
(4)pk(1 p)n
-

f(k) = P(Binom(n , p) = () = -

nMRMun

=A
f(() = P(Poisson(x) = 1)

111111111 ....

P(E = K) = 0 since I is continuous

instead

p(d[[b)
or

pl0IX)
"this is the sof of I

P([[X) = 1 -

p(I]X)
= 1-P(no people appear

during [0 , X])
= 1 -

p(# = 0 , where # is the

Poisson (XX) RV)
= 1 -
(xx(0f(x)
0!
x
=Gf(x) =d(l-e-x)
f(x) =
 Lecture : April ?"a
I -E xp(0)

e
=
then f(()

where X = avg visitors unit intervened

per

1
0 :
avg rate =

properties of Exp(0)

Se
f(()) , 20
1 .
=
is pdf

O if (0

>
- f((()20 for all X

(8f(())d)) 1
>
- =

-
C

2 .

M
=
ESI] =
0
*
xf(x)dx
((t)
g
=

88

e-Xod
=
.
3 Var(I) = 02 do

f(x)d
f(X- 5)
>
var() =

O
-
 Lecture : April ?"a
Example :

mean rate : 30 students

prob of wait times 3 min

# # of students
#
= waiting time

# = Poisson (0 3) .

I student/min
# =
Exp(2) 2 min before student
3)
@gIe-WI dN
> ...
PCE =

normal RV
1 inflection points
+ o &
M M o
-

.
2 about
symmettric the M

3 . area under curve =

1

gf() -=
4 . f(x))0

3 .
f(x) + 0 as (()) >-
@

6 . max f((() =
f(M)
X =
IR
video lecture
Lecture : April 4th
obj : investigate variable of interest

sol : take representative samples

sample mean

30042
Sty St 3
= =
reduces the SE
larger samples

Bias(E) = [E] -

O
[(E) = 0 unbiased

sample mean for large n

My M =
Ox =
@
M

example 1
bank has so accounts

M = 1 2009 sd 300$
= n = 50

1 , 200 242 . 3
200
= SE O
:
My

1 230-2 , 200 1
M
2 : 2 =
,

42 43
.

P (2) 1 .

18)ed2 PL1

P(211 18).
= 0 1198
.

P : 11 9 %
.
 Lecture : April 4th
probability that I card balance above $1250

:
2

X = 1250

M
= 1200

0 = 300

2 = 0 167
.

cumulative probability
2 = 0 167
.

P(2[0 167)v0 5663. .

↑ (x) 1250) = 1 -
P(280 167)
.
= 0 4335
.

Pr 43 35 % .