lecture 14
DESCRIPTION
Lecture 14TRANSCRIPT
UNR, MATH/STAT 352, Spring 2007
Binomial(n,p)
!( ) (1 ) (1 )
( )! !k n k k n kn n
P X k p p p pk n k k
( )E X np 2 (1 )X np p
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
0.1
0.2
0.3
0.4
0.5
Number of succsses
db
ino
m(q
, 1, 0
.5)
Binomial(1,0.5)
Number of successes within 1symmetric Bernoulli trialcan only be 0 or 1. These
possibilities have equal chances.
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
0.1
0.2
0.3
0.4
0.5
Number of succsses
db
ino
m(q
, 2, 0
.5)
Binomial(2,0.5)
Number of successes within 2symmetric Bernoulli trials
can only be 0, 1 or 2. The possibility to have exactly 1 success
is larger than that of having 0 or 2.
A fair game should result in a tie.
Then why do people play fair games?
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
0.2
0.4
0.6
0.8
Number of succsses
db
ino
m(q
, 2, 0
.1)
Binomial(2,0.1)
Most likely there will be no successes
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
0.2
0.4
0.6
0.8
Number of succsses
db
ino
m(q
, 2, 0
.9)
Binomial(2,0.9)
Most likely there will be only successes
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
00
.05
0.1
00
.15
0.2
00
.25
0.3
00
.35
Number of succsses
db
ino
m(q
, 15
, 0.1
)
Binomial(15,0.1)
Unimodal (mode = 1)
Right-skewed
Concentrated around 1.5
(E = np = 1.5)
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
00
.05
0.1
00
.15
0.2
00
.25
Number of succsses
db
ino
m(q
, 19
, 0.1
)
Binomial(19,0.1)
Unimodal (mode = 1-2)
Right-skewed
Concentrated around 1.5
UNR, MATH/STAT 352, Spring 2007
0 5 10 15 20 25 30
0.0
00
.02
0.0
40
.06
0.0
80
.10
0.1
2
Number of succsses
db
ino
m(q
, 10
0, 0
.1)
Binomial(100,0.1)
Unimodal (mode = 10)
Symmetric? (E = np = 10)
Concentrated around 10
P(10+3) < P(10-3)
UNR, MATH/STAT 352, Spring 2007
70 75 80 85 90 95 100
0.0
00
.02
0.0
40
.06
0.0
80
.10
0.1
2
Number of succsses
db
ino
m(q
, 10
0, 0
.9)
Binomial(100,0.9)
Unimodal (mode = 90)
Symmetric? (E = np = 90)
Concentrated around 90
UNR, MATH/STAT 352, Spring 2007
0 20 40 60 80 100
0.0
00
.02
0.0
40
.06
0.0
80
.10
0.1
2
Number of succsses
db
ino
m(q
, 10
0, 0
.1)
Bin(100,0.9)Bin(100,0.1)
Only a small fraction of possible outcomes has not negligible P (i.e. only small part can be seen
in experiment)
P is very small (not 0!) here
UNR, MATH/STAT 352, Spring 2007
0 1 2 3 4 5 6
0.0
00
.05
0.1
00
.15
0.2
00
.25
0.3
0
Number of succsses
db
ino
m(q
, 6, 0
.5)
Binomial(6,0.5)
Here all possibleoutcomes have reasonable
probabilities
UNR, MATH/STAT 352, Spring 2007
Poisson()
( )!
k
P X k ek
( )E X 2X
UNR, MATH/STAT 352, Spring 2007
0 2 4 6 8 10
0.0
0.1
0.2
0.3
Number of succsses
dp
ois
(q, 1
)Poisson(1)
0 2 4 6 8 10
0.0
0.1
0.2
0.3
Number of succsses
db
ino
m(q
, 10
00
, 1/1
00
0)
Binomial(1000,.001]
I’ve seen this already!
UNR, MATH/STAT 352, Spring 2007
n np 0p+ +
Binomial(n,p) Poisson()UNR, MATH/STAT 352, Spring 2007
0 10 20 30 40 50
0.0
00
.02
0.0
40
.06
Number of succsses
dp
ois
(q, 3
0)
0 10 20 30 40 50
0.0
00
.02
0.0
40
.06
Number of succsses
dp
ois
(q, 3
0)
0 10 20 30 40 50
0.0
00
.02
0.0
40
.06
Number of succsses
dp
ois
(q, 3
0)
Poisson
Binomial
Normal
Poisson(30) Binomial(1000,.03) N(30,sqrt(30))
UNR, MATH/STAT 352, Spring 2007
If n is large (n > 100), p is small (p < 0.05), andboth np and n(1-p) are not small (say >10) then
B(n,p)~P(np)~N(np, np(1-p))
UNR, MATH/STAT 352, Spring 2007
Normal densities
Value of random variable
De
nsi
ty
-10 -8 -6 -4 -2 0 2 4 6 8 10
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Mean= 7, Std=1
Mean= 5, Std=1Mean= 0, Std=1
Mean= 0, Std=0.6
Mean= 0, Std=3
UNR, MATH/STAT 352, Spring 2007
2400 2450 2500 2550 2600
0.00
00.
006
0.01
2Probability density function
Value of random variable
Pro
babi
lity
Lower-tail probability
Interval probability
Upper-tail probability
2400 2450 2500 2550 2600
0.0
0.4
0.8
Cumulative distribution function
Value of random variable
Pro
babi
lity
UNR, MATH/STAT 352, Spring 2007
UNR, MATH/STAT 352, Spring 2007
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
2 games. P(lose or win < $10)= 1
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
4 games. P(lose or win < $10)= 1
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
6 games. P(lose or win < $10)= 1
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
10 games. P(lose or win < $10)= 1
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
100 games. P(lose or win < $10)= 0.728
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
1000 games. P(lose or win < $10)= 0.272
Net payoff, in $
Pro
babi
lity
of p
ayof
f
UNR, MATH/STAT 352, Spring 2007
12Payoff 2 , where Bin( , )
(Why this formula? See lecture notes.)
X N X N
-100 -50 0 50 100
0.0
0.1
0.2
0.3
0.4
0.5
10000 games. P(lose or win < $10)= 0.087
Net payoff, in $
Pro
babi
lity
of p
ayof
f