Probability

1
  We want to set up the mathematical framework for
dealing with probability. Loosely speaking, the
probability of an event is a numerical measure
(between 0 and 1 inclusive) of how likely it is that
the event will happen.
 By an experiment we simply mean an activity with
Event and an observable outcome.
Probabilities  The sample space of an experiment is the set of
possible outcomes of that experiment.
 A particular outcome is called a sample point.
 A collection of possible outcomes is called an
event.

2
  Flipping a coin. The sample space
S1 = {H,T}
And the particular outcome “the coin lands head” is
the sample point H
 Throwing a die. The sample space is
S2 = {1,2,3,4,5,6}
Examples The outcome “the top face shows a three is the
sample point 3. The event A1, that the die shows an
even number is the subset A1 = {2,4,6} of the sample
space. And the event A2 that the dies shows a number
larger than 3 is the subset A2={4,5,6}

3
  A set is a collection of objects of interest, the
members of the set are called elements.
If A and B are sets then:
1. Union (A U B)
2. Intersection (A INTER B)
Reminder:
3. Subset
Sets and
4. Empty set
Notation
5. Complement (A BARRE)
With respect to a given sample space, S, the
complement of A, Ac, is the set of elements which are
in S but not in A;

4
 Recall that events are just subsets of the sample
space. So we can naturally rewrite statements
about events in set theory notation

The event A or B occurs =>

The event A and B occurs =>
Example
The event A does not occur
=> COMPLEMENT A

A collection of sets or events are said to be disjoint or

mutually exclusive if they have no elements in
common. In such a case at most one of the events can
occur each time the experiment is performed.
Intersection is an empty space nothing in commun

5
 In some settings where the sample space is finite
(fini) , it is natural to assume in advance that each
possible outcome of the experiment is as likely as any
other to occur. (Coin and die but not Protein
sequence)
Equally In this case we can calculate the probability of an
event as P(A) = |A|/|S|
Likely
Where |A| denotes the number of sample points in
Probability the event A.
In this “equally likely” setting, calculating
probabilities reduces to counting the number of
elements in the events in question
Numb of elements in A /numb of elements in b

6
  If an event can occur in N mutually exclusive and
equally likely ways, and if m of these possess a trait
Definition A, the probability of the occurrence of A is equal to
m/N =P(A)

7
 1. The die flipping example A1= even and A2>3
P(A1) = 3/6=1/2 and P(A2) = 3/6 =1/2
P(A1 int A2) = P(face showing is even and larger than
3) = 2/6 =1/3

Examples 2. Throw a fair die twice. What is the probability of

the event E that the sum of the faces showing on
each toss is 9?
Sample space = ?
Event?

8
 What is the probability of the event F that a hand in
bridge (i.e. 13 cards) contains 5spades, 4 hearts, 3
diamonds, and 1 club

Need to define my event

Probability= |p| / |s|

9
  In many setting the sample space is infinite or all
possible outcomes of the experiment are not equally
likely. We still wish to associate probabilities with events
of interest. Such assignment of probabilities to events
must satisfy certain conditions or axioms to avoid
inconsistencies
Probability in  The three axioms which must be satisfied are

more general 1. 0≤P(A) ≤1 for any event A

settings 2. P(S) = 1
3. If A1, A2, .. are mutually exclusive then

10
Properties

11
  For any two events A and B
Addition
Rule

 Extension: P(A or B or C)??

12
  Single nucleotide polymorphisms (SNPs) are
nucleotide positions in a genome which exhibit
variation amongst individuals in a species. In some
studies in humans, SNPs are discovered in
European populations. Suppose that of such SNPs,
Example 70% also show variation in African population, 80%
also in Asian population and 60% in both African
and Asian population.
 Suppose one SNP is chosen at random, what is the
probability that it is variable on either the African or
the Asian population?

13
  Sometimes we will have partial information about
the outcome of the experiment. In general this
partial information will change probabilities.
Conditional Definition
Probability  Provided P(B) >0, the conditional probability of A
given B is written as

14
Properties of
Conditional
Probability

15
  Two events A and B are said to be independent if

 If A and B are independent

Independenc
e of Events
 It is important to be aware that the terms
independent and mutually exclusive do not mean
the same thing.

16
Frequency of Family History of Mood Disorder by Age Group among
Bipolar Subjects
Family History of Mood Early = 18 (E) Later > 18 (L) Total
Disorders
Negative (A) 28 35 63
Bipolar Disorder (B) 19 38 57
Unipolar (C) 41 44 85
Unipolar and Bipolar (D) 53 60 113
Total 141 177 318

1. Suppose we pick a person at random from this sample. What

is the probability that this person will be 18 years old or
younger?
2. Suppose we pick a subject at random from the 318 and find
that he is 18 years or younger. What is the probability that
this subject will be one who has no family history of mood
disorders? (Conditional Probability)
3. What is the probability that a person picked at random will
be Early and will be a person who has no family history of
mood disorders? (Multiplication Rule, Joint Probability)
17
  In a population of a particular species of snail,
individuals, exhibit different forms. It is known that
45% have a pink background coloring while 55%
have a yellow background coloring. In addition,
30% of individuals are striped, and 20% of the
population are pink and stripped.
Example  Is the presence or absence of striping independent
of background color?
 Given that a snail is pink, what is the probability
that it will have stripes?

P(A inter B) = P(A) * P(B)

18
  In many contexts there are different ways an event can
happen, and to calculate the probability of the event one
needs to add up the probabilities associated with each of
the different ways it can happen. This can be expressed
mathematically as follows.
Partition  Suppose the events E1, .. En form a partition of the
sample space S (by which we mean that exactly one of
Rule the events Ei must happen) then for any A

19
  Mendelian Segregation
At a particular locus in human, there are two alleles,
A and B, and it is known that the population
frequencies of the genotypes AA, AB and BB are
0.49, 0.42, and 0.09 respectively.
An AA man has a child with a woman whose
Example genotype is unknown
1. What is the probability that the child will have
genotype AB?
2. Suppose the child is genotyped and found to be
AB. What is the probability that the mother is
BB?

20
21
  The partition rule works out the probability of an event by
breaking it up according to the different ways it might
have happened.
 It might happen that we want to know the opposite. We
are told the result and asked which of the events of the
Bayes partition occurred
Theorem

22
23
24
25
26
27
  In the health sciences field, a widely used
application of probability laws and concepts is
found in the evaluation of screening tests and
diagnostic criteria. (i.e. presence/absence of
Screening disease,…)
Tests  Also of interest is the likelihood of positive and
negative test result.
 However, we must be aware of the fact that they
are not always infallible.

28
 1. A false positive (Type I error ) results when a test
indicates a positive status when the true status
is negative.
Definitions 2. A false negative (Type II error) results when a
test indicates a negative status when the true
status is positive.

29
 In summary, the following questions must be
answered in order to evaluate the usefulness of test
results and symptom status in determining whether
or not a subject has some disease:
1. Given that a subject has the disease, what is the
probability of a positive test result?
Screening 2. Given that a subject does not have the disease,
Tests what is the probability of a negative test result?
3. Given a positive screening test, what is the
probability that the subject has the disease?
4. Given a negative screening test result, what is
the probability that the subject does not have
the disease?

30
 Disease
Test Result Present (D) Absent (Dc) Total
Positive (T) a b a+b
Negative c d c+d
(Tc)
Total a+c b+d n

Screening • a = True Positive b = False Positive

Test • c = False Negative d = True
Negative

• P(T|D) = a/(a+c) = sensitivity of the test.

The sensitivity of a test is the probability of a positive test
result given the presence of the disease

• P(Tc|Dc) = d/(b+d) = specificity of the test.

The specificity of a test is the probability of a negative
test result given the absence of the disease 31
 Disease
Test Result Present (D) Absent (Dc) Total
Positive (T) a b a+b
Negative c d c+d
(Tc)
Total a+c b+d n
Screening
Test • P(D|T) = a/(a+b) = predictive value positive = number
of TP/number of P.  Precision
The predictive value positive of a screening test is the
probability that a subject has the disease given that the
subject has a positive screening test result.

• P(Dc|Tc) = d/(c+d) = predictive value negative.

The predictive value negative of a screening test is the
probability that a subject does not have the disease, given
that the subject has a negative screening test result.
32
  Predictive Value Positive (Precision) and predictive
value negative can only be estimated using data
from a cross-sectional study or other population-
Screening based study in which prevalence estimates may be
Test obtained.
 In contrast, specificity and sensitivity can be
estimated from case-control studies.

33
  Estimates of the predictive value positive and predictive
value negative of a test may be obtained from a
knowledge of a test’s sensitivity and specificity and the
Screening probability of the relevant disease in the general
population (prevalence).
Test
 For this we use the Bayes' theorem.

34
Screening
These two events are mutually
Test exclusive

Using the multiplication

rule

35
 Predictive value positive

Screening
Test Predictive value negative

36
  The Accuracy of a test is the proportion of true
results (both true positives and true negatives) in
the population.

Screening
 Accuracy = (a+d)/(a+b+c+d)
Test
 Accuracy = (sensitivity)(prevalence) + (specificity)
(1-prevalence)

37
 Alzheimer’s Diagnosis?
Test Result Yes (D) No (Dc) Total
Positive (T) 436 5 441
Negative 14 495 509
(Tc)
Total 450 500 950

Example
1. Sensitivity?
2. Specificity?
3. Compute the predictive value positive of the
test.
Evans et al estimated that 11.3 percent of the
US population aged 65 and over have
Alzheimer’s disease.
38