Escolar Documentos
Profissional Documentos
Cultura Documentos
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
Brian Caffo
Department of Biostatistics
Johns Hopkins Bloomberg School of Public Health
Johns Hopkins University
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Table of contents
Brian Caffo
1 Table of contents
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
3 Binomial trials
Properties
ML estimate of
Properties
ML estimate of
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
P(X = x) = p x (1 p)1x
The mean of a Bernoulli random variable is p and the variance is p(1 p)
If we let X be a Bernoulli random variable, it is typical to call X = 1 as a
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
is
n
Y
The Bernoulli
distribution
Properties
ML estimate of
xi
(1 p)n
xi
i=1
Binomial trials
The normal
distribution
p xi (1 p)1xi = p
xi /n is
The normal
distribution
Properties
ML estimate of
0.4
Binomial trials
0.2
The Bernoulli
distribution
0.0
Table of
contents
likelihood
Brian Caffo
0.6
0.8
1.0
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
0.0
0.2
0.4
0.6
p
0.8
1.0
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Binomial trials
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
The binomial random variables are obtained as the sum of iid Bernoulli trials
In specific, let X1 , . . . , Xn be iid Bernoulli(p); then X =
random variable
The binomial mass function is
Properties
ML estimate of
P(X = x) =
for x = 0, . . . , n
n
x
p x (1 p)nx
Pn
i=1 Xi
is a binomial
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Properties
ML estimate of
=
n!
x!(n x)!
Binomial trials
The normal
distribution
n
x
n
0
=
n
n
=1
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Consider the probability of getting 6 heads out of 10 coin flips from a coin
Binomial trials
p 6 (1 p)4
The normal
distribution
Properties
ML estimate of
There are
10
6
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Example
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
Suppose a friend has 8 children, 7 of which are girls and none are twins
If each gender has an independent 50% probability for each birth, whats the
0.8
1.0
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
The Bernoulli
distribution
Binomial trials
0.6
0.4
Table of
contents
Likelihood
Brian Caffo
0.0
Properties
ML estimate of
0.2
The normal
distribution
0.0
0.2
0.4
0.6
p
0.8
1.0
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
2 /2 2
Binomial trials
The normal
distribution
Properties
ML estimate of
normal distribution
The standard normal density function is labeled
Standard normal RVs are often labeled Z
0.4
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
0.3
Brian Caffo
Binomial trials
0.2
The Bernoulli
distribution
density
Table of
contents
0.0
Properties
ML estimate of
0.1
The normal
distribution
0
z
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
X = + Z N(, 2 )
The non-standard normal density is
{(x )/}/
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
Approximately 68%, 95% and 99% of the normal density lies within 1, 2 and 3
standard deviations from the mean, respectively
1.28, 1.645, 1.96 and 2.33 are the 10th , 5th , 2.5th and 1st percentiles
of the standard normal distribution respectively
By symmetry, 1.28, 1.645, 1.96 and 2.33 are the 90th , 95th , 97.5th and 99th
percentiles of the standard normal distribution respectively
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
Question
What is the 95th percentile of a N(, 2 ) distribution?
We want the point x0 so that P(X x0 ) = .95
X
x0
P(X x0 ) = P
x0
= P Z
= .95
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
Therefore
x0
= 1.645
or x0 = + 1.645
In general x0 = + z0 where z0 is the appropriate standard normal quantile
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Question
What is the probability that a N(, 2 ) RV is 2 standard deviations above the
mean?
We want to know
Binomial trials
The normal
distribution
Properties
ML estimate of
P(X > + 2) = P
X
+ 2
>
= P(Z 2)
2.5%
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Other properties
1
The normal distribution is symmetric and peaked about its mean (therefore
the mean, median and mode are all equal)
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Question
If Xi are iid N(, 2 ) with a known variance, what is the likelihood for ?
Brian Caffo
Table of
contents
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
n
Y
L() =
(2 2 )1/2 exp (xi )2 /2 2
i=1
(
exp
n
X
)
(xi )2 /2 2
i=1
(
= exp
n
X
i=1
xi2 /2 2
n
X
)
2
Xi / n /2
i=1
exp n
x / 2 n2 /2 2
Later we will discuss methods for handling the unknown variance
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
The Bernoulli
distribution
Question
If Xi are iid N(, 2 ), with known variance whats the ML estimate of ?
We calculated the likelihood for on the previous page, the log likelihood is
n
x / 2 n2 /2 2
Binomial trials
The normal
distribution
Properties
ML estimate of
n
x / 2 n/ 2 = 0
This yields that x is the ml estimate of
Since this doesnt depend on it is also the ML estimate with unknown
Mathematical
Biostatistics
Bootcamp:
Lecture 7,
Some
Common
Distributions
Brian Caffo
Table of
contents
Pn
i=1 (Xi
The Bernoulli
distribution
Binomial trials
The normal
distribution
Properties
ML estimate of
)2
X
n
Which is the biased version of the sample variance
The ML estimate of is simply the square root of this estimate
To do likelihood inference, the bivariate likelihood of (, ) is difficult to
visualize
Later, we will discuss methods for constructing likelihoods for one parameter
at a time