Professional Documents
Culture Documents
Probability 020423
Probability 020423
139
Chapter9
Probability
9.1. PROBABILITY
Number of cases favorable to the
9.1.1. Concept or P(A)= 0ccurrence of the event
Computation ofthe chance of an Total number of mutually exclusivc
Crentis
called
probability. Hence, occurrence
probability
of a
is thecertain and exhaustive cases
random happening of an event and study
of
"Probability of rain
"Probability of
tomorrow uncertainty,
is only 35%".
e.g., So the total number of trials favorable to the event (A)
winning lottery is
a l out of
and divided by the total number of ways in which the event
40,406,353". (A) can happen.
Before conducting an experiment, the
the outcome is unpredictable. therefore absolute
it is certainty
of
Number of cases favorable to the
9.1.2. Meaning and Definition of Probability against the occurrence of the event
Probability
Probability is the
The number of cases against he
mode of expression of knowledge or 0ccuTence of the event
belief whether the event has occurred or will occur. Thus, m The number of cases favorable to the
in the theory of probability this concept has been given an occurrence of the event
exact mathematical significance for being used widely in
Such areas of study as mathematics, statistics, finance, 9.1.3.
gambling, science, and philosophy, so that conclusions can Characteristics of Probability
De drawn regarding the probability of potential events and Function P(A)
1) The limit of probability of an event A is between 0
Ihe mechanics of complex systems being hidden.
and 1, i.e. 0sP(A)<I.
ACOrding to Levin, We live in a world in which we are 2) When probability of two events A and B which are
Unable to forecast the future with complete certainty. Our mutually exclusive, is combined, the probability is the
ed to cope with uncertainty leads us to the study and use sum of individual probabilities of A and B, i.e.. P (A
of probability theory." UB)will be P (A) + P (B).
3) Two events which are equivalent, are
assigned the
According to American Heritage Dictionary,
is the branch of mathematics that studies the
same probability.
"Prikelobabiihoodlity of OCcurrence of random events in order to 9.1.4. Applications of Probability
Pehct the behavior of a defined system."
99
1) It plays a great role in making predictions.
2) It is used to solve scientific investigations as well as
General Rule for Probability as follows:
day-to-day problems.
The
1f an probability of an event (say A) is defined
fails n times, then the
3) It is very useful when one wants to predict the
uncertainties of betting and chances of success.
event happens mtimes and it A) It is extensively used in business and
probability (p) of event Ais defined as: 5) It is the basis of the law of 'statisticaleconomics.
regularity' and
PA)= mmtn law of inertia of a large number'.
(Business Statistics and Analytics) ,
140 (Unit-IV) MBA First Semester
AKTU
9.1.5. Uses of in 9.2.2. Objective Probability
Probability The pobability of the 0ccurrcnce of an event is entircly
Decision-Making
Following arc the busincss
situations
probability to take a decision:
wlhere maagets ase
bascd on the analysis whcre cach measure
documcntcd obscrvation. in placc of a
depends o
subjective
estimate.
A noe accurale way of determining probabilities is by
D
Investnent Problen: AcOmpally's manager has (Wo
of objcctive probabilities
in
diflerent projects (with difterent initial Wily Conrast to the
costs) Tor observations bascd on subjcctive measures, like
investment. The decision taken by the manager
bascd on the choice that the outconne of is estimates. For example, the objcctive probability i. personal
which is determined for tossing a coin 100 times and
dependent on the level of demand. cach observation that the outcomc of the coin after landi. recording
2) Introducing a New Product: While
new product there arises a problem introducing
in deciding
a will be *heads". In statistical analysis, it is very imporant
that each observation must be an independent event
about the introduction of the product. The decision without manipulations. The lesser the observation beine
maker is not sure whether the product
would be biased, the lesser will be the cnd probability.
acceptable or not.
Hence after conducting test, Objective probabilities are of two types:
different regions the results beingmarketing
in three 1) Classical Probability, and
idea of introducing a new product contradictory, the 2) Empirical Probability.
dropped? Thus, it is necessary to answer should be
this 9.2.2.1. Classical Probability
question: what would be the chances of success of
introducing a new product? Classical Probability is the first approach to the theory of
3) Stocking probability.
Decisions: The demand of a perishable
commodity is unknown to its dealer in advance. According to Laplace, "Probability is the 'ratio of the
if the commodity is not sold by the end Thus number of favorable cases to the total number of equally
of the day it
will be spoiled. Therefore, he is not sure
about the likely cases."
demand pattern and stock amount (that how much The fundamentai assumption to this theory is that the
commodity is to be stored in advance). various outcomes of an event are 'equally likely'. Thus the
4) The Individual Investor:
Maximisation of returns is probability of happening of these events is also equal. In
the motive of investors who are this theory, the probability of happening of an
engaged in buying event is
determined prior to the happening of the event. Therefore
and selling of equities. The behaviour of equities /
security prices is uncertain and it depends upon these probabilities are known as priori
various factors. In this situation, managers take classical theory is suitable for determinationprobability. The
of those events that are possible of probability
decisions on the basis of their prediction about future through games of chance.
prices of the securities. This prediction helps the where various outcomes are equally likely to
investor to take a decision as to which securities the tossing of a dice the happen. In
six possible ways. Thus happening
of the event can be in
he/she should select for the investment. the
event is described as 'p' and probability of success of
the probability of the failre
of an event is termed as q' where
9.2. PROBABILITY there is no thirdevent.
P= Number of favorable cases
THEORY Total number of equally likely
Let there be 'x ways for the
cases
9.2.1. Introduction "y' ways for its failure and oceurrence of an event
they are equally likely to OC
The branch of mathematics that is concerned with random
then the probability of the
phenomena analysis is known as probability theory. The happening of the event
Occurrence of a random event cannot be determincd and it denoted by p. These
can be any single outcome from various pOssible unitary, theoretical or probabilities
mathematical
are also known as
outcomes. The actual outcome can be determined only by probability of the occurrence of the probability.
event and qPe
is the
chance. There are different definitions of probability
which are shown in figure below:
probability of its notoccurrence.
p:
X +y and q=
Types of Probability X+ y
So, X
p+q= y
Objective Probability (x + y) (x+ y)
Subjective X+y
Probability Theory Therefore, p + q=1, or 1
-q=p, or
Classical Probability
The l-p4
Empirical
Probability
probability of
between the numberoccurrence cvent E is the
of cases inof itsan favour
ratio
and the total
number of cases (equally likely).
Probability (Chapter9) |41
PE)=
n(E) Number of cases favourable to event E event is derived from the past experiences
or from the
the output of
n(S) Total numberof cases relative frequencies of success in the past. If
a machine is 10% of unacceptable articles in the past then
isthe number of cases favourable to the event E, y articles
the number of cascs favourable to the event E'. the odds inis the relative frequency for those unacceptable
Would be 10% of the total items.
favour of E are x:y and odds against of E
are y:x.
Thus, the estimation of the relative frequency should be on
In this case. the basis of cxtcnsive readings in the past. The accuracy of
P(E)=X P(E)= the result is directly proportional to the large number of
the
X +y X+ y past readings. The probabilities that are calculated on
basis of past experience are called posterior
.. P (E) + P(E) = 1 probabilities, and have relative frequency approach. This
is in contrast to the priori probabilities that are calculated
Therefore, probab1lity of happening + probability of non through classical approach.
happening = 1 and 0s P(E)S1, thus,
of P(E)=] and the m
maximum value
minimum value of P(E) =0. The priori probabilities are applicd in the calculation of the
games of chance whereas posterior probabilities are
Probabilities are expressed either as ratio, fraction or applied in the problems of economic and social
percentage, such as 1or 0.5 or 50%. phenomena where priori probabilities are not constant.
The nature of the priori probabilities is deductible and is
theory-based rather than being based on evidence,
They arc presumed to be equally likely and the probability experience and experimentation. Posterior probabilities
of getting a 6 on a single throw of a dice would be /6. which are also known as empirical probabilities, and are
Limitations of Classical Approach based on the past experience as well as on conducted
) The limitation of this definition is that, it is restricted experiments.
only to the games of chancc and the problens other
than the game of chance cannot be explained. Thus the posterior probability or cmpirical probability (P)
of an event is given by,
) This method is not applicable when it is not
to calculate the total number of cases. possible No. of times an event is observed
P=
3) This method is not applicable in the case where No.of trails experiment is conducted
outcomes of a random cxperiment are not equally likely.
4) The further subdivision of the possible If, out of 1000 items produced by a machine in past, 60
outcome of
experiments int0 mutually exclusive, exhausive ::nd were found to be defective, the probability of a
cqually likely isalnnost difficult. defective article to be produced by this machine would
60
9.2.2.2. Relative be or 0.06.
Frequency Approach/ 1000
Empirical Probability
The scientific study of measuring uncertainty is known as Lmitations of Relative Frequently Theory of Probability
I) In conducting Jarge number of trials of the
'probabiliy'. Probabilities are empiically deterinined experiment
there may not exist given homogeneous and identical
when their nunerical values are based upon asample or a
conditions of the experimnent.
census of data which gives a distribution of events.
2) No natter iow large is the value of n,
the relative
According to Von Mises, "If an experiment is performed frequency m/n, will not attain a unique value.
repeatedly under essentially homogeneous and identical 3) The defined
Conditions, then the limiting value of the ratio of the probability (P) can never be obtained
nùmber of times the event occurs to the number of trials. practically. An attempt can be made at a nearby
as the nunmber of trials becomes indefinitely large. iscalled estimate of Pby adequately increasingn.
the probability of happening of the event. it is being 9.2.3.
aSsumed that the limit is finite and unique". Subjective Approach to
Symbolically, if A is the name of un event, f is the
Probabilityand
The classical
equency with which that event occurred. and n is the
empirical approach of probability is
objective in description. Whercas the subjective
sample size, then: probability of an event is considercd to be the access to
scope of
one's certainty of a particular event lo occur.
P(A)= considers an event that the student "A' Hence, if one
wi pass the
examination, then its probability cannot be estimated by
Tiori probabilities are not the busis of the relative cither of the objective approaches that have
quency approach to the theory of probability. There are above. The passing and faiiing of a student been discussed
are the events
nous situations where it is impossible to have equaliy which are not equally likely cases and if
these cases were
ACIV events which form the besis of the classicai theory equai!y likely, thern by applying the classical
probability. Thus. the probability of the happening of an probability we could have derived the probability approach of
to be
142 (Unit-IV) MBA First Semester(Business Statistics and Analytics) AKTU
1/2. Thus, in this case the outcomes (n, = 6). The toss of coin will have two possible
under uniform conditions. experiment cannot be repeated that theroe
With the help of empirical outcomes (n, = 2), The counting rule suggests
approach it is not possible to comment be (6) x (2) = 12 distinct experimental outcomes.
upon the
probability of this event. Hence comparatively, subjective
approach is useful in such This is also clear from the sample space of the experiment
the probability of an eventcases because in this approach
is the
of tossingtwo coins given above as S ={(1,.,T), (1, H), (2,
representation of the
degree of faith and belief of a rational T), (2, H), (3, T), (3, H), (4, T), (4, H), (3, T), (5, H), (6. T
occurrence that particular event that depends
of person in the (6, H)}, i.e., in an experiment of throwing a dice and tossing
judgment, personal point of view, etc. upon his acoin, there are six possible outcomes (tigure 9.1),
Considering this approach, the probability of an event is 9.3.2. Counting Rules for Permutations
different for each person. Due to this it is
subjective probability where the probability known as Apermutation is a particular sequence of objects (or simple
affected by the of an event is events) where the order of selection foms subsets nr
is also known aspersonal bias of its estimator. Therefore, it arrangements that are considered unique or distinctive
Personalised
the presumption that Theory of Probability on
any decision is the reflection According to the counting rules for permutations the possible
personality of its decision-maker along with of the number of experimental outcomes whenr items are to he
important subjective elements that assign a those selected from a set of r items in a specific order. The sanme :
an event. Let us probability to
assume that have data relating to the items selected in a different order would be considered a
price of a share for the last 3 we different experimental outcome. The total number of
that out of 2000 quotations years and further supposing permutations of nitems, taking rat a time, is given by,
there was price rise on 500 relating to this share price, n
(3,6), (4, 5), (5, 4), (6. 3), and hence P(B) =364 Pla jack) =2
52 13
|46(Unit-|V) MBA First Scmeslcr (Business Statistics and Analytics) AKTU
combined/simultaneous
Occurrence of several events, the theorem of multiplication
probability is used. The description of this P(AnB)=xm m m,
P(AnB)=P(A). P(B/A)
Multiplication Theorem Note: In the case of three events A, B, C, P(A n
BO)=
P(A). PB/A). ie., the probability of occurrence of
Multiplication Theorem
on Probabilities for
A.B
A, B and C is equal to the
Indenendent Events Multiplication Theorem
on Probabilities for probability of B given that Aprobability of A times the
has occurred, times the
Dependent Events probability of C given that bothA and B have occurred.
1) Multiplication Theorem Example 6: A coin is tossed
three times in
on
Independent Events: If two events A and B are Probabilities for succession. What is the probability that all the three
are heads?
independent, then the probability of simultaneous
occurrence of both the events is equal to the product of Solution: Let,
their individual properties. Thus, P(A B) =P(A).PB) A denote the event that
first toss
Proof: Out of n, possible cases let m, denote the event that second toss is ais "a head",B
favorable for the occurrence of the event cases be denote the event that third toss is "a head",
C
A. head".
P(A) n, Now the result of the first
toss
n, influence the result of the second onedoes
and
in no way
hence A, B.
Out of D, possible cases, let m, cases be Care all independent events.
for the occurrence of the event B favourable P(A) =P(B) =P(C) = Moreover,
P(B) "2 .:. P (that all the
three are heads) =
n, =
P(A).P(B).P(C) as all events arePAnBOC)
Each of n, possible cases can be
associated with each independen
of the n, possible cases.
Therefore the total number of possible cases for
occurrence of the event A' and B' is n x the nT.
Example 7: A person is known to hit a target in 4
Similarly each of the m, favourable cases can be of 5shots, where as 0u!
associated with each of the m, favourable cases. So 3out of 4 shots. Findanother person is known to hit
hit at all when the probability that the target 1S
the total number of favourable cases for the they both try.
event 'A'
and B' is mx m2. Solution: Let A be the event that the first
P(AnB)= m, m, m, m, = P(A). P(B) the target, person n
Note
i) The theorem can be extended to three or more
PA)=and P(A»)==
Let Bbe the
independent events. If A, B, C.... be independent event that another person hits the target,
events, then P(A BnC...) = P(A).P(B).P(C)... P(B) =and P(B*) =4
Probab1lity (Chapter 9) |47
target[Independent events]
= P (that the
Using the reduced sample space E, we find that:
.. P(X) is hit at all) 440 11
P(ME) = 600
=1-P(X) =1- 1 19 15
20 20
Let n(A) denote the number of elements in any set A.
Using this notation, we can write:
9.6. CONDITIONAL n(E M) n(En M)/ n(S) n(EnM)
PROBABILITY P(ME) = n(E) n(E)/n(S) P(E)
Where, P(En M) and P(E) are found from the original
Generally, conditional probability analysis involves sample space S. To verify this result, note that:
estimating the probability of occurence of a particular 600 2 440 22
event (A) as conditional upon the occurrence of a P(E)= and P(E OM) =
narticular event (B). This probability is usually 900 900 45
P(A/B) and is read "the probability of A given B."written
as
Hence, P(MIE) = 22/45 22x3 11
, as before.
Or 2/3 45x2 15
For (wo events A and B, then the conditional
of A(given B) is that the probability of probability Example 8: A pair of dice is rolled. If the sum of 8 has
that B has occurred) is given by: occurring A (given appeared, find the probability that one of the dice shows 3.
P(AB)= P(AnB) Provided P(B) > 0. Solution: The equi-probable sample space consisting of
P(B) 36 sample points.
The event A =The sumn of the scores is 8 has five sample
Similarly, the conditional probability of B, given A, points (2, 6), (3, 5), (4, 4), (5, 3), (6, 2) and it has reduccd
denoted by P(B) is defincd by:
sample space.
P(B|A) = P(AnB)
P(A)
Provided P(A)>0. P(A) =
n(A)
n(S) 36
For example, consider the experiment of
eight sided dice. There can be eight possible throwing
outcomes
an
in
Under the assumption that A has happened, the event B =
one of the dice shows 3 has only two
this experiment. These eight possible outcomes will form a sample points, that is
sample space. So, the sample space will be, (AnB)= [(3, 5), (5, 3)].
S= {1,2, 3, 4, 5, 6,7, 8} .:. P(AOB) = n(AnB) 2 1
Let A be the event of getting an odd number, B is the n(S) 36 18
event getting at least 7. Also using the formula, we get p(RA) P(AnB)
Then, A=(1,3, 5, 71, B =(7, 8) and AnB=(7) 1/18 136 2 P(A)
4 5/36 18 5
:. P(A)=*= P(B) = and P(A B) =
2 Example 9: An box contains 3 white and 5
draws of one ball each are made red balls. Two
P(B0A) P(A^B)_1/8_1!: What is the probability that one is redwithout replacement.
P(A) 1/2 S 1 4
Solution: Probability of drawing a and the other white?
red ball in the first
P(A|B) = P(ANB)_1/8
P(B) 1/4 8 1
draw. P(A)= 5
For example, suppose that our sample space S is the Probabilityof drawing a white ball in
the second
Population of adults in small town who have completed
given that the first draw has given a red balldraw
3 or
requirements for a college degree. We shall categorise P (since only 7 balls are left in the
them according to gender and emplovment status. The data 7 bag and three
are given in table 9.1. white balls are still there).
(Business Statistics and
148 (Unit-IV) MBA First Semester Analytics)
Posterior
)AKTU
However, the Bayesian, or the
Probability of the combined event
P(AB)= P(A n B)= P(A)×
probabilities
always conditional probabilities which are calculated arefor
P(B/A) = 53 15 every events as follows:
87 56 1) Mutually Excusive Events: If there exists
But it could also happen that in the combination with
first draw a white ball Occurrence of an event Ein
was drawn then, Probability of
drawing a white b¡ll in the the mutually exclusive events Ej, Ez ...E, then one of
first draw, P(A) = and P(E, )P(E /E,)
8 PERE)=
Probability of drawing a red ball in the second draw given
that the first draw gave a
white ball or P Where k=1,2, ....n
The combined probability of the
two events is, 2) Mutually Exclusive and Exhaustive Events: Thacs
P (AB) or P (AnB) = P 15 are two mutually exclusive and exhaustive evente A
(A)x P =(BIA) and A, and the occurrence of the event B is
7 56
Now any one of the two combination with A, and A, then for event Athe
situations
ball first or we draw a white ball (when we draw a red
first), would conditional probability, given the event Bis as under.
conditions of the problem. These two events aresatisfy the P(A,)P(B /lA,)
exclusive. So the probability that any one of mutually P(A, /B) =
happens is the sum of the two probabilities i.e., the two P(A)P(B/A)+P(A,)P(B/A,)'
Required probability = 15 15 15 P(A, /B) = P(A,)P(B /lA,)
56 56 28
P(A,)P(B/lA,)+ P(A,)P(B/A,)
9.7 BAYE'S THEOREM 9.7.1. Need of Baye's Theorem
There are various names assigned to this Baye's theorem is applied in the case where the following
like, posterior probability, revised probability, conditions exist in the problems:
probability,
inverse probability. This was introduced by Thomasor 1) The sample space is further divided into a
set of
Baye (an English mathematician) in his famous work mutually exclusive events (A, Az... Ag}.
on Bayesian Decision Theory. This theory states 2) Within the sample space,
about event B exists, lor which
the finding of the probability of an
event by P(B) > 0.
the available sample information. considering 3) Computation of a conditional
P (Ag| B) is the detailed probability of the form:
According to Baye's Theorem uncertainty can be 4) Either one of the two sets of
objective.
measured. On the basis of probability theory, this theorem below. probabilities is described
establishes a rule for refining a hypothesis by allotting i)
additional evidence and circumstantial information which P(A, n B) for each A7,
results in a number, representing the degree of probability ii) P(A;) and P(B |A,) for
that hypothesis is true. For example, for an event A., a each A:
chosen sample of 3 defective itemns out of 100 can be used 9.7.2.
to estimate the probability for an event B that a machine is
1) The Applications of Baye's Theorem
not working properly. theorem stipulates inmultiplying the preceus
distribution by the likelihood
to attain posterior function and directing
The Bayesian probability is based on the conditional
probability, where A, and A, are the events that are 2) Baye's theorem being
distribution.
mutually exclusive and exhaustive. Whereas B is asimple fomal is valid to all commou
event which intersects with each of the A vents as interpretations of probability. Hence
Bayesian interpretation. probabilitiesaccording to
displayed in the Venn diagrams as shown in figure 9.3. are subject
logical coherent degree of faith, or adegree offaithin
the given proposition pertaining to the body of well
A A specified information.
B
Example 10: Abag Acontains 3 white and 4black balls
and a bag Bcontains 5 white and 3 black bals. One ballis
Figure 9.3: Bayesian Probability drawn at random from one of the bags and it is foundtto be
This is known as posterior probability as it is calculated black. Find the probability that it was drawn fromthe bagB.
after the information isrecorded. It is the revised probability Solution: Let, X;= The event of ball being drawnfrom
as it is derived by the revision of the prior probabilities bag A,
under the accumulated additional information. Further, it is X,= The event of ball being drawn from bag
also kncwn as inverse probability, i.., determining the
B and
probability of a probability.
X=The event of ba!l being black