Scribd
Upload
59 views44 pages

Chap 5 Prob Distributions I

This document provides an overview of several common probability distributions: 1) The binomial distribution describes the number of successes in a fixed number of independent yes/no experiments, and is characterized by the number of trials (n) and probability of success (p). 2) The Poisson distribution models the number of rare, independent events occurring in a fixed time period or space, and is characterized by the average rate of occurrences (λ). 3) The exponential distribution describes the time between events in a Poisson process, and has rate parameter λ. 4) The normal/Gaussian distribution is symmetric and bell-shaped, characterized by a mean (μ) and standard deviation (σ). It arises from the sum

Uploaded by

MbusoThabethe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
59 views44 pages

Chap 5 Prob Distributions I

This document provides an overview of several common probability distributions: 1) The binomial distribution describes the number of successes in a fixed number of independent yes/no experiments, and is characterized by the number of trials (n) and probability of success (p). 2) The Poisson distribution models the number of rare, independent events occurring in a fixed time period or space, and is characterized by the average rate of occurrences (λ). 3) The exponential distribution describes the time between events in a Poisson process, and has rate parameter λ. 4) The normal/Gaussian distribution is symmetric and bell-shaped, characterized by a mean (μ) and standard deviation (σ). It arises from the sum

Uploaded by

MbusoThabethe
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 44

Chapter 5

Probability distributions I

Binomial, Poisson, Exponential and


Normal Distributions
BINOMIAL DISTRIBUTION
• Also called Bernoulli distribution
• Motivating example:
• Example 1A: A door-to-door salesperson calls
on 6 clients per session. Each client makes
their purchasing decision independently of the
others, with probability 0.2 of purchasing the
product. What are the probabilities that 0, 1,
2, 3, 4, 5 or 6 clients purchase the product?
Motivating example
• Experiment:
– Salesperson visits 6 clients
– Trail = a single visit
• Two possible outcomes: Success (S), Failure (F)
• Probability of Success does not change during
the experiment: P(S) = 0.2 = p
• Trails are independent of each other
• Random variable X: Number of successes
Motivating example
• Determine the probability of two successes:
• How many different ways can the salesperson
achieve 2 S’s and 4 F’s.
• E.g: SSFFFF, SFSFFF, SFFSFF,…, FFFFSS
• Combination:
• Probability:
– P(SSFFFF) = P(S) x P(S) x P(F) x P(F) x P(F) x P(F)
– =p2(1-p)4
Motivating example
• Hence,
–
BINOMIAL DISTRIBUTION
• In a binomial process, we have n independent trials, each trial
has two outcomes, success or failure, and P[success] = p for
all trials. Let the random variable X be the number of
successes in n trials.
• Then X has the binomial distribution, and P[X = x] is given by
the probability mass function:

• X ~ Binomial(n, p)
Examples
• Calculate the binomial function in Excel.
• Draw bar charts similar to the ones on pp 116-
117.
• Example 5C, p118
POISSON DISTRIBUTION
• Classical example: Radioactive decay
• Characteristics:
– Random process that occurs at constant rate
• E.g:
– Number of alpha particles measured per minute
– Number of customers arriving at the ATM machine
per hour
• Rate per unit of time is
POISSON DISTRIBUTION
• Time or space
– Poisson process can refer to events that occur
randomly over a spatial or temporal dimension.
– Number of faults in fiber optic cable per km
– Number of misprints per page in a book
– Number of diamonds per cubic meter of ore
RADIOACTIVE DECAY
• From experiment by Dr James Parks, U. Tennessee
RADIOACTIVE DECAY - DATA
RADIOACTIVE DECAY - DATA

λ = 3.2084 counts per 0.2 seconds


POISSON DISTRIBUTION
POISSON DISTRIBUTION
• Draw bar charts of Poisson distribution in
Excel, for λ = 3 and λ = 8
• Example 6A
• Example 8C
Example 9C
Beer cans are randomly tossed alongside the national road, with an
average frequency 3.2 per km.
a) What is the probability of seeing no beercans over a 5 km stretch?
b) What is the probability of seeing at least one beercan in 200 m?
c) Determine the values of x and y in the following statement: “40%
of 1 km sections have x or fewer beercans, while 5% have more
than y.”
DERIVATION OF POISSON FORMULA

• Events occur randomly at rate λ per time interval.


• Partition the unit time into n equal parts

• λ events occur per unit time interval. Hence the


probability that it occur in part of the partition is:
DERIVATION OF POISSON FORMULA

• Assume that n is no large, that there are only


at most one event per internal 1/n of the unit
interval.
• Now we have the requirements for the
Binomial distribution:
– n independent trials
– Two outcomes: success or failure
– p = probability of success, does not change from
trial to trail
DERIVATION OF POISSON FORMULA

• Hence, if X is the number of events per unit


time:
• X ~ B(n, λ/n)
DERIVATION OF POISSON FORMULA
DERIVATION OF POISSON FORMULA
EXPONENTIAL DISTRIBUTION
• Related to the Poisson distribution
• Same experiment “Poisson process”:
– Process with constant average rate λ, events occur
continuosly and independently etc
– Random variable X is now waiting time between
events
– Sample space S = {x|x ≥ 0}
• It has the key property of being memoryless.
Applications
• In real-world scenarios, constant rate (or probability per unit time) is
rarely satisfied. E.g. the rate of incoming phone calls differs according to
the time of day.
• But if we focus on a time interval during which the rate is roughly
constant, such as from 2 to 4 p.m. during work days, the exponential
distribution can a good approximate model for the time until the next
phone call arrives.
• In queuing theory, the service times of agents in a system (e.g. how long it
takes for a bank teller etc. to serve a customer) are often modeled as
exponentially distributed variables.
• Similar caveats apply to the following examples which yield approximately
exponentially distributed variables:
– The time until a radioactive particle decays, or the time between clicks of a geiger
counter
– The time until default (on payment to company debt holders) in reduced form credit risk
modeling
EXPONENTIAL DISTRIBUTION

• Density function, not mass function


Example 10A, p122
NORMAL OR GAUSSIAN DISTRIBUTION
• Discovered by Carl F Gauss and independently by
Laplace.
• Also discovered by Abraham de Moivre, but it did
not become well known in his time.
• Importance of normal distribution stems from
the Central Limit Theorem.
• CLT states that if a random variable X is the sum
of a large number of random increments, then X
follows the normal distribution.
• Show Excel simulation
NORMAL DISTRIBUTION

• Symmetric distribution
• Bell-shaped
• Two parameters:
– Mean: μ
– Standard deviation: σ
STANDARD NORMAL DISTRIBUTION

• By doing a transform, the N.D. can be


transformed into the Standard Normal
Distribution.
Example 15A, p126
Example 15A, p126
Example 17B, p129
A t-shirt manufacturer knows that the chest
measurements of his customers are normally
distributed with mean 92 cm and standard
deviation 5 cm. He makes his t-shirts in four sizes
— S ( fit size range 80–87 cm), M (to fit 87–94), L
(to fit 94–101) and XL (to fit 101–108). What
proportion of customers fit into each size t-shirt?
Example 17B, p129
Example 20C, p131
• A soft-drink vending machine is set to discharge an
average of 215 ml of cooldrink per cup. The amount
discharged is normally distributed with standard
deviation 10 ml.
• (a) If 225 ml cups are used, what proportion of cups
overflow?
• (b) What is the probability that a cup contains at least
200 mℓ of cooldrink?
• (c) What size cups ought to be used if it is desirable
that only 2% of cups overflow?
SUMS AND DIFFERENCES OF
INDEPENDENT NORMAL DISTRIBUTIONS
• What happens when we add two or more
random variables together, all coming from
normal distributions?

• let
• then
• Where and
•
Example 22C
Plastic caps seal the ends of the tube into which your
degree certificate is placed when you graduate. Suppose
the tubes have a mean diameter of 24.0mm and a standard
deviation of 0.15mm, and that the plastic caps have a
mean diameter of23.8 mm and a standard deviation of
0.11mm. If the diameter of the cap is 0.10 mm or more
larger than that of the tube, the cap cannot be squashed
into the tube, and if the diamater of the cap is 0.45 mm or
more smaller than that of the tube, it will not seal the
tube, but will just keep falling out. If a tube and and plastic
cap are selected at random, what are the probabilities of:
(a) the cap being too large for the tube, and
(b) the cap falling out of the tube?
MULTIPLYING A NORMAL RANDOM
VARIABLE WITH A CONSTANT
• Example: Celsius to Fahrenheit
• Also consider the case where you’d like to
convert from degrees Celsius (X) to degrees
Fahrenheit (Y ). It is known that the
relationship is given by:

• So in general, if then
PERCENTAGE POINTS OF THE STANDARD
NORMAL DISTRIBUTION
• Often, instead of wanting to find Pr[Z > z ] for some given value of z , we
are given a probability p and need to find the value of z that makes the
equation Pr[Z > z] = p true. It is convenient to use the notation z(p) to
denote the value of z which provides the solution to this equation, and to
describe it as the 100p% point of the distribution.
MULTIPLYING A NORMAL RANDOM
VARIABLE WITH A CONSTANT
• Example: Celsius to Fahrenheit
• Also consider the case where you’d like to
convert from degrees Celsius (X) to degrees
Fahrenheit (Y ). It is known that the
relationship is given by:

• So in general, if then
• So:
=
Finally:

You might also like