Probability Distributions

Содержание

Слайд 2

Random Variable

A random variable x takes on a defined set of values

Random Variable A random variable x takes on a defined set of
with different probabilities.
For example, if you roll a die, the outcome is random (not fixed) and there are 6 possible outcomes, each of which occur with probability one-sixth.
For example, if you poll people about their voting preferences, the percentage of the sample that responds “Yes on Proposition 100” is a also a random variable (the percentage will be slightly differently every time you poll).
Roughly, probability is how frequently we expect different outcomes to occur if we repeat the experiment over and over (“frequentist” view)

Слайд 3

Random variables can be discrete or continuous

Discrete random variables have a countable

Random variables can be discrete or continuous Discrete random variables have a
number of outcomes
Examples: Dead/alive, treatment/placebo, dice, counts, etc.
Continuous random variables have an infinite continuum of possible values.
Examples: blood pressure, weight, the speed of a car, the real numbers from 1 to 6.

Слайд 4

Probability functions

A probability function maps the possible values of x against their

Probability functions A probability function maps the possible values of x against
respective probabilities of occurrence, p(x)
p(x) is a number from 0 to 1.0.
The area under a probability function is always 1.

Слайд 5

Discrete example: roll of a die

Discrete example: roll of a die

Слайд 6

Probability mass function (pmf)

Probability mass function (pmf)

Слайд 7

Cumulative distribution function (CDF)

Cumulative distribution function (CDF)

Слайд 8

Cumulative distribution function

Cumulative distribution function

Слайд 9

Examples

1. What’s the probability that you roll a 3 or less?
P(x≤3)=1/2
2.

Examples 1. What’s the probability that you roll a 3 or less?
What’s the probability that you roll a 5 or higher?
P(x≥5) = 1 – P(x≤4) = 1-2/3 = 1/3

Слайд 10

Practice Problem

Which of the following are probability functions?
a.      f(x)=.25 for x=9,10,11,12
b.      f(x)=

Practice Problem Which of the following are probability functions? a. f(x)=.25 for
(3-x)/2 for x=1,2,3,4
c. f(x)= (x2+x+1)/25 for x=0,1,2,3

Слайд 11

Answer (a)

a.      f(x)=.25 for x=9,10,11,12

1.0

Answer (a) a. f(x)=.25 for x=9,10,11,12 1.0

Слайд 12

Answer (b)

b.      f(x)= (3-x)/2 for x=1,2,3,4

Answer (b) b. f(x)= (3-x)/2 for x=1,2,3,4

Слайд 13

Answer (c)

c. f(x)= (x2+x+1)/25 for x=0,1,2,3

Answer (c) c. f(x)= (x2+x+1)/25 for x=0,1,2,3

Слайд 14

Practice Problem:

The number of ships to arrive at a harbor on any

Practice Problem: The number of ships to arrive at a harbor on
given day is a random variable represented by x. The probability distribution for x is:

Find the probability that on a given day:
a.    exactly 14 ships arrive
b.    At least 12 ships arrive
c.    At most 11 ships arrive

 p(x=14)= .1

p(x≥12)= (.2 + .1 +.1) = .4

p(x≤11)= (.4 +.2) = .6

Слайд 15

Practice Problem:

You are lecturing to a group of 1000 students. You ask

Practice Problem: You are lecturing to a group of 1000 students. You
them to each randomly pick an integer between 1 and 10. Assuming, their picks are truly random:
What’s your best guess for how many students picked the number 9?
Since p(x=9) = 1/10, we’d expect about 1/10th of the 1000 students to pick 9. 100 students.
What percentage of the students would you expect picked a number less than or equal to 6?
Since p(x≤ 6) = 1/10 + 1/10 + 1/10 + 1/10 + 1/10 + 1/10 =.6 60%

Слайд 16

Important discrete distributions in epidemiology…

Binomial
Yes/no outcomes (dead/alive, treated/untreated, smoker/non-smoker, sick/well, etc.)
Poisson
Counts (e.g.,

Important discrete distributions in epidemiology… Binomial Yes/no outcomes (dead/alive, treated/untreated, smoker/non-smoker, sick/well,
how many cases of disease in a given area)

Слайд 17

Continuous case

The probability function that accompanies a continuous random variable is a

Continuous case The probability function that accompanies a continuous random variable is
continuous mathematical function that integrates to 1.
The probabilities associated with continuous functions are just areas under the curve (integrals!).
Probabilities are given for a range of values, rather than a particular value (e.g., the probability of getting a math SAT score between 700 and 800 is 2%).

Слайд 18

Continuous case

For example, recall the negative exponential function (in probability, this is

Continuous case For example, recall the negative exponential function (in probability, this
called an “exponential distribution”):

This function integrates to 1:

Слайд 19

Continuous case: “probability density function” (pdf)

The probability that x is any exact

Continuous case: “probability density function” (pdf) The probability that x is any
particular value (such as 1.9976) is 0; we can only assign probabilities to possible ranges of x.

Слайд 20

For example, the probability of x falling within 1 to 2:

For example, the probability of x falling within 1 to 2:

Слайд 21

Cumulative distribution function

As in the discrete case, we can specify the “cumulative

Cumulative distribution function As in the discrete case, we can specify the
distribution function” (CDF):
The CDF here = P(x≤A)=

Слайд 23

Example 2: Uniform distribution

The uniform distribution: all values are equally likely
The uniform

Example 2: Uniform distribution The uniform distribution: all values are equally likely
distribution:
f(x)= 1 , for 1≥ x ≥0

Слайд 24

Example: Uniform distribution

 What’s the probability that x is between ¼ and ½?

Example: Uniform distribution What’s the probability that x is between ¼ and

P(½ ≥x≥ ¼ )= ¼

Слайд 25

Practice Problem

4. Suppose that survival drops off rapidly in the year

Practice Problem 4. Suppose that survival drops off rapidly in the year
following diagnosis of a certain type of advanced cancer. Suppose that the length of survival (or time-to-death) is a random variable that approximately follows an exponential distribution with parameter 2 (makes it a steeper drop off):

What’s the probability that a person who is diagnosed with this illness survives a year?

Слайд 26

Answer

The probability of dying within 1 year can be calculated using

Answer The probability of dying within 1 year can be calculated using
the cumulative distribution function:

The chance of surviving past 1 year is: P(x≥1) = 1 – P(x≤1)

 
Cumulative distribution function is:

Слайд 27

Expected Value and Variance

All probability distributions are characterized by an expected value

Expected Value and Variance All probability distributions are characterized by an expected
and a variance (standard deviation squared).

Слайд 28

For example, bell-curve (normal) distribution:

For example, bell-curve (normal) distribution:

Слайд 29

Expected value, or mean

If we understand the underlying probability function of a

Expected value, or mean If we understand the underlying probability function of
certain phenomenon, then we can make informed decisions based on how we expect x to behave on-average over the long-run…(so called “frequentist” theory of probability).
Expected value is just the weighted average or mean (µ) of random variable x. Imagine placing the masses p(x) at the points X on a beam; the balance point of the beam is the expected value of x.

Слайд 30

Example: expected value

Recall the following probability distribution of ship arrivals:

Example: expected value Recall the following probability distribution of ship arrivals:

Слайд 31

Expected value, formally

Discrete case:

Continuous case:

Expected value, formally Discrete case: Continuous case:

Слайд 32

Empirical Mean is a special case of Expected Value…
Sample mean, for a

Empirical Mean is a special case of Expected Value… Sample mean, for
sample of n subjects: =

Слайд 33

Expected value, formally

Discrete case:

Continuous case:

Expected value, formally Discrete case: Continuous case:

Слайд 34

Extension to continuous case: uniform distribution

x

p(x)

1

1

Extension to continuous case: uniform distribution x p(x) 1 1

Слайд 35

Symbol Interlude

E(X) = µ
these symbols are used interchangeably

Symbol Interlude E(X) = µ these symbols are used interchangeably

Слайд 36

Expected Value

Expected value is an extremely useful concept for good decision-making!

Expected Value Expected value is an extremely useful concept for good decision-making!

Слайд 37

Example: the lottery

The Lottery (also known as a tax on people who

Example: the lottery The Lottery (also known as a tax on people
are bad at math…)
A certain lottery works by picking 6 numbers from 1 to 49. It costs $1.00 to play the lottery, and if you win, you win $2 million after taxes.
If you play the lottery once, what are your expected winnings or losses?

Слайд 38

Lottery

Calculate the probability of winning in 1 try:

The probability function (note, sums

Lottery Calculate the probability of winning in 1 try: The probability function (note, sums to 1.0):
to 1.0):

Слайд 39

Expected Value

The probability function

Expected Value

E(X) = P(win)*$2,000,000 + P(lose)*-$1.00
= 2.0 x

Expected Value The probability function Expected Value E(X) = P(win)*$2,000,000 + P(lose)*-$1.00
106 * 7.2 x 10-8+ .999999928 (-1) = .144 - .999999928 = -$.86

Negative expected value is never good!
You shouldn’t play if you expect to lose money!

Слайд 40

Expected Value

If you play the lottery every week for 10 years, what

Expected Value If you play the lottery every week for 10 years,
are your expected winnings or losses?
520 x (-.86) = -$447.20

Слайд 41

Gambling (or how casinos can afford to give so many free drinks…)

A

Gambling (or how casinos can afford to give so many free drinks…)
roulette wheel has the numbers 1 through 36, as well as 0 and 00. If you bet $1 that an odd number comes up, you win or lose $1 according to whether or not that event occurs. If random variable X denotes your net gain, X=1 with probability 18/38 and X= -1 with probability 20/38.
E(X) = 1(18/38) – 1 (20/38) = -$.053
On average, the casino wins (and the player loses) 5 cents per game.
The casino rakes in even more if the stakes are higher:
E(X) = 10(18/38) – 10 (20/38) = -$.53
If the cost is $10 per game, the casino wins an average of 53 cents per game. If 10,000 games are played in a night, that’s a cool $5300.

Слайд 42

**A few notes about Expected Value as a mathematical operator:

If c= a

**A few notes about Expected Value as a mathematical operator: If c=
constant number (i.e., not a variable) and X and Y are any random variables…
E(c) = c
E(cX)=cE(X)
E(c + X)=c + E(X)
E(X+Y)= E(X) + E(Y)

Слайд 43

E(c) = c

E(c) = c
Example: If you cash in soda

E(c) = c E(c) = c Example: If you cash in soda
cans in CA, you always get 5 cents per can.
Therefore, there’s no randomness. You always expect to (and do) get 5 cents.

Слайд 44

E(cX)=cE(X)


E(cX)=cE(X)
Example: If the casino charges $10 per game instead of $1, then

E(cX)=cE(X) E(cX)=cE(X) Example: If the casino charges $10 per game instead of
the casino expects to make 10 times as much on average from the game (See roulette example above!)

Слайд 45

E(c + X)=c + E(X)
E(c + X)=c + E(X)
Example, if the casino

E(c + X)=c + E(X) E(c + X)=c + E(X) Example, if
throws in a free drink worth exactly $5.00 every time you play a game, you always expect to (and do) gain an extra $5.00 regardless of the outcome of the game.

Слайд 46

E(X+Y)= E(X) + E(Y)

E(X+Y)= E(X) + E(Y)
Example: If you play

E(X+Y)= E(X) + E(Y) E(X+Y)= E(X) + E(Y) Example: If you play
the lottery twice, you expect to lose: -$.86 + -$.86.
NOTE: This works even if X and Y are dependent!! Does not require independence!! Proof left for later…

Слайд 47

Practice Problem

If a disease is fairly rare and the antibody test is

Practice Problem If a disease is fairly rare and the antibody test
fairly expensive, in a resource-poor region, one strategy is to take half of the serum from each sample and pool it with n other halved samples, and test the pooled lot. If the pooled lot is negative, this saves n-1 tests. If it’s positive, then you go back and test each sample individually, requiring n+1 tests total.
Suppose a particular disease has a prevalence of 10% in a third-world population and you have 500 blood samples to screen. If you pool 20 samples at a time (25 lots), how many tests do you expect to have to run (assuming the test is perfect!)? 
What if you pool only 10 samples at a time?
5 samples at a time?

Слайд 48

Answer (a)

a. Suppose a particular disease has a prevalence of 10% in

Answer (a) a. Suppose a particular disease has a prevalence of 10%
a third-world population and you have 500 blood samples to screen. If you pool 20 samples at a time (25 lots), how many tests do you expect to have to run (assuming the test is perfect!)?
Let X = a random variable that is the number of tests you have to run per lot:
E(X) = P(pooled lot is negative)(1) + P(pooled lot is positive) (21)
E(X) = (.90)20 (1) + [1-.9020] (21) = 12.2% (1) + 87.8% (21) = 18.56
E(total number of tests) = 25*18.56 = 464

Слайд 49

Answer (b)

b. What if you pool only 10 samples at a

Answer (b) b. What if you pool only 10 samples at a
time?
E(X) = (.90)10 (1) + [1-.9010] (11) = 35% (1) + 65% (11) = 7.5 average per lot
50 lots * 7.5 = 375

Слайд 50

Answer (c)

c. 5 samples at a time?
E(X) = (.90)5 (1) + [1-.905]

Answer (c) c. 5 samples at a time? E(X) = (.90)5 (1)
(6) = 59% (1) + 41% (6) = 3.05 average per lot
100 lots * 3.05 = 305

Слайд 51

Practice Problem

If X is a random integer between 1 and

Practice Problem If X is a random integer between 1 and 10,
10, what’s the expected value of X?

Слайд 52

Answer

If X is a random integer between 1 and 10, what’s the

Answer If X is a random integer between 1 and 10, what’s
expected value of X?

Слайд 53

Expected value isn’t everything though…

Take the show “Deal or No Deal”
Everyone know

Expected value isn’t everything though… Take the show “Deal or No Deal”
the rules?
Let’s say you are down to two cases left. $1 and $400,000. The banker offers you $200,000.
So, Deal or No Deal?

Слайд 54

Deal or No Deal…

This could really be represented as a probability distribution

Deal or No Deal… This could really be represented as a probability
and a non-random variable:

Слайд 55

Expected value doesn’t help…

Expected value doesn’t help…

Слайд 56

How to decide?

Variance!
If you take the deal, the variance/standard deviation

How to decide? Variance! If you take the deal, the variance/standard deviation
is 0.
If you don’t take the deal, what is average deviation from the mean?
What’s your gut guess?

Слайд 57

Variance/standard deviation

“The average (expected) squared distance (or deviation) from the mean”

**We square

Variance/standard deviation “The average (expected) squared distance (or deviation) from the mean”
because squaring has better properties than absolute value. Take square root to get back linear average distance from the mean (=”standard deviation”).

Слайд 58

Variance, formally

Discrete case:

Continuous case:

Variance, formally Discrete case: Continuous case:

Слайд 59

Similarity to empirical variance
The variance of a sample: s2 =

Similarity to empirical variance The variance of a sample: s2 =

Слайд 60

Symbol Interlude

Var(X) = σ2
these symbols are used interchangeably

Symbol Interlude Var(X) = σ2 these symbols are used interchangeably

Слайд 61

Variance: Deal or No Deal

Now you examine your personal risk tolerance…

Variance: Deal or No Deal Now you examine your personal risk tolerance…

Слайд 62

Practice Problem

A roulette wheel has the numbers 1 through 36, as well

Practice Problem A roulette wheel has the numbers 1 through 36, as
as 0 and 00. If you bet $1.00 that an odd number comes up, you win or lose $1.00 according to whether or not that event occurs. If X denotes your net gain, X=1 with probability 18/38 and X= -1 with probability 20/38.
We already calculated the mean to be = -$.053. What’s the variance of X?

Слайд 63

Answer

Standard deviation is $.99. Interpretation: On average, you’re either 1 dollar above

Answer Standard deviation is $.99. Interpretation: On average, you’re either 1 dollar
or 1 dollar below the mean, which is just under zero. Makes sense!

Слайд 64

Handy calculation formula!

Handy calculation formula (if you ever need to calculate by

Handy calculation formula! Handy calculation formula (if you ever need to calculate by hand!):
hand!):

Слайд 65

Var(x) = E(x-μ)2 = E(x2) – [E(x)]2 (your calculation formula!)

Proofs (optional!):
E(x-μ)2

Var(x) = E(x-μ)2 = E(x2) – [E(x)]2 (your calculation formula!) Proofs (optional!):
= E(x2–2μx + μ2) remember “FOIL”?!
=E(x2) – E(2μx) +E(μ2) Use rules of expected value:E(X+Y)= E(X) + E(Y)
= E(x2) – 2μE(x) +μ2 E(c) = c
= E(x2) – 2μμ +μ2 E(x) = μ
= E(x2) – μ2
= E(x2) – [E(x)]2
OR, equivalently:
E(x-μ)2 =

Слайд 66

For example, what’s the variance and standard deviation of the roll of

For example, what’s the variance and standard deviation of the roll of a die?
a die?

Слайд 67

**A few notes about Variance as a mathematical operator:

If c= a constant

**A few notes about Variance as a mathematical operator: If c= a
number (i.e., not a variable) and X and Y are random variables, then
Var(c) = 0
Var (c+X)= Var(X)  
Var(cX)= c2Var(X)
Var(X+Y)= Var(X) + Var(Y) ONLY IF X and Y are independent!!!!
{Var(X+Y)= Var(X) + Var(Y)+2Cov(X,Y) IF X and Y are not independent}

Слайд 68

Var(c) = 0

Var(c) = 0
Constants don’t vary!

Var(c) = 0 Var(c) = 0 Constants don’t vary!

Слайд 69

Var (c+X)= Var(X)

Var (c+X)= Var(X)
Adding a constant to every instance of a

Var (c+X)= Var(X) Var (c+X)= Var(X) Adding a constant to every instance
random variable doesn’t change the variability. It just shifts the whole distribution by c. If everybody grew 5 inches suddenly, the variability in the population would still be the same.

Слайд 70

Var (c+X)= Var(X)

Var (c+X)= Var(X)
Adding a constant to every instance of a

Var (c+X)= Var(X) Var (c+X)= Var(X) Adding a constant to every instance
random variable doesn’t change the variability. It just shifts the whole distribution by c. If everybody grew 5 inches suddenly, the variability in the population would still be the same.

Слайд 71

Var(cX)= c2Var(X)

Var(cX)= c2Var(X)
Multiplying each instance of the random variable by c makes

Var(cX)= c2Var(X) Var(cX)= c2Var(X) Multiplying each instance of the random variable by
it c-times as wide of a distribution, which corresponds to c2 as much variance (deviation squared). For example, if everyone suddenly became twice as tall, there’d be twice the deviation and 4 times the variance in heights in the population.

Слайд 72

Var(X+Y)= Var(X) + Var(Y)

Var(X+Y)= Var(X) + Var(Y) ONLY IF X and Y

Var(X+Y)= Var(X) + Var(Y) Var(X+Y)= Var(X) + Var(Y) ONLY IF X and
are independent!!!!!!!!
With two random variables, you have more opportunity for variation, unless they vary together (are dependent, or have covariance): Var(X+Y)= Var(X) + Var(Y) + 2Cov(X, Y)

Слайд 73

Example of Var(X+Y)= Var(X) + Var(Y): TPMT

TPMT metabolizes the drugs 6- mercaptopurine,

Example of Var(X+Y)= Var(X) + Var(Y): TPMT TPMT metabolizes the drugs 6-
azathioprine, and 6-thioguanine (chemotherapy drugs)
People with TPMT-/ TPMT+ have reduced levels of activity (10% prevalence)
People with TPMT-/ TPMT- have no TPMT activity (prevalence 0.3%).
They cannot metabolize 6- mercaptopurine, azathioprine, and 6-thioguanine, and risk bone marrow toxicity if given these drugs.

Слайд 74

TPMT activity by genotype

Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt 2):601-5

TPMT activity by genotype Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt 2):601-5

Слайд 75

TPMT activity by genotype

Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt 2):601-5

The

TPMT activity by genotype Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt
variability in TPMT activity is much higher in wild-types than heterozygotes.

Слайд 76

TPMT activity by genotype

Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt 2):601-5

There

TPMT activity by genotype Weinshilboum R. Drug Metab Dispos. 2001 Apr;29(4 Pt
is variability in expression from each wild-type allele. With two copies of the good gene present, there’s “twice as much” variability.

No variability in expression here, since there’s no working gene.

Слайд 77

Practice Problem

Find the variance and standard deviation for the number of ships

Practice Problem Find the variance and standard deviation for the number of
to arrive at the harbor (recall that the mean is 11.3).

Слайд 78

Answer: variance and std dev

Interpretation: On an average day, we expect 11.3

Answer: variance and std dev Interpretation: On an average day, we expect
ships to arrive in the harbor, plus or minus 1.35. This gives you a feel for what would be considered a usual day!

Слайд 79

Practice Problem

You toss a coin 100 times. What’s the expected number of

Practice Problem You toss a coin 100 times. What’s the expected number
heads? What’s the variance of the number of heads?

Слайд 80

Answer: expected value

Intuitively, we’d probably all agree that we expect around 50

Answer: expected value Intuitively, we’d probably all agree that we expect around
heads, right?
Another way to show this?
Think of tossing 1 coin. E(X=number of heads) = (1) P(heads) + (0)P(tails)
∴E(X=number of heads) = 1(.5) + 0 = .5
 If we do this 100 times, we’re looking for the sum of 100 tosses, where we assign 1 for a heads and 0 for a tails. (these are 100 “independent, identically distributed (i.i.d)” events)
E(X1 +X2 +X3 +X4 +X5 …..+X100) = E(X1) + E(X2) + E(X3)+ E(X4)+ E(X5) …..+ E(X100) =
100 E(X1) = 50

Слайд 81

Answer: variance

What’s the variability, though? More tricky. But, again, we could do

Answer: variance What’s the variability, though? More tricky. But, again, we could
this for 1 coin and then use our rules of variance.
Think of tossing 1 coin.
E(X2=number of heads squared) = 12 P(heads) + 02 P(tails)
∴E(X2) = 1(.5) + 0 = .5
Var(X) = .5 - .52 = .5 - .25 = .25
Then, using our rule: Var(X+Y)= Var(X) + Var(Y) (coin tosses are independent!)
Var(X1 +X2 +X3 +X4 +X5 …..+X100) = Var(X1) + Var(X2) + Var(X3)+ Var(X4)+ Var(X5) …..+ Var(X100) =
100 Var(X1) = 100 (.25) = 25
SD(X)=5

Interpretation: When we toss a coin 100 times, we expect to get 50 heads plus or minus 5.

Слайд 82

Or use computer simulation…

Flip coins virtually!
Flip a virtual coin 100 times; count

Or use computer simulation… Flip coins virtually! Flip a virtual coin 100
the number of heads.
Repeat this over and over again a large number of times (we’ll try 30,000 repeats!)
Plot the 30,000 results.

Слайд 83

Coin tosses…

Mean = 50
Std. dev = 5
Follows a normal distribution
∴95% of the

Coin tosses… Mean = 50 Std. dev = 5 Follows a normal
time, we get between 40 and 60 heads…

Слайд 84

Covariance: joint probability

The covariance measures the strength of the linear relationship between

Covariance: joint probability The covariance measures the strength of the linear relationship
two variables
The covariance:

Слайд 85

The Sample Covariance

The sample covariance:

The Sample Covariance The sample covariance:
Имя файла: Probability-Distributions.pptx
Количество просмотров: 41
Количество скачиваний: 0