The Poisson Distribution
Videos
1. Introduction
2. Adapting the rate
3. Example 1
4. Sum of distributions
Statistics 1 Review
What is a discrete random variable?
What are the properties of a probability distribution?
Name the 2 discrete distributions studied in S1 and the conditions for each
distribution.
How is the random variable defined in each distribution?
Lesson 1
Poisson Distribution
Used in modelling the number of events over a unit of time or space
Examples of situations that can be modelled using Poisson distribution are:
• the number of automobiles arriving at a toll gate in 5-minute interval.
• the number of emergency calls received by an ambulance control in an hour
• the number of arrivals at a car wash in one hour
• the number of network failures per day
• the number of file server virus infection at a data centre during a 24-hour period
• the number of flaws per metre of cable
• the number of cars passing a stretch of road per day
• the number of weeds per square metre in a field.
Conditions for a Poisson model (state these conditions in the context of the problem)
1) Events occur singly in a given interval of time or space
2) Events occur independently and randomly in a given interval of time or space
3) λ, the mean(average) number of occurrences in an interval is constant
The variable X is the number of occurrences in the given interval.
If the above conditions are satisfied, X is said to follow a Poisson distribution, written:
X ̴ Po (λ), where
𝜆𝑥
P (X =x) = 𝑒 −𝜆 for x = 0, 1, 2, 3 ,…
𝑥!
If X ̴ 𝑃𝑜 (λ)
E (X) = λ and Var(X) = λ
Question 1
A student finds that the average number of bacteria in 10 ml of pond water from a
particular pond is 4. Find the probability that in a 10 ml sample
𝜆𝑥
Assume that the number of bacteria follows a Poisson distribution, P (X =x) = 𝑒 −𝜆
𝑥!
a) There are exactly five bacteria
b) There are no bacteria
c) There are fewer than three bacteria
Answers
a) 0.156 b) 0.0183 c) 0.238
What happens if the interval changes?
Question 2
Suppose the interval changes. Suppose instead of the number of bacteria in 10ml of water
we are looking at the number of bacteria in 1 ml of water.
Then, what is the probability that:
a) There are exactly two bacteria in 1ml of water
Answer:
a) 0.0536
Question 3
(Take care to specify the interval being considered)
On average, the school photocopier breaks down eight times during the school week
(Monday-Friday). Assuming that the number of breakdowns can be modelled by a Poisson
distribution, find the probability that it breaks down:
a) Five times in a given week
b) Once on Monday
c) Eight times in a fortnight
Answers
a) 0.0916 b) 0.323 c) 0.0120
Question 4
X follows a Poisson distribution with standard deviation 1.5.
Find P (X ≥ 3)
Answer:
0.391
Homework
Worksheet Questions 3 and 5
Additional Practice
Exercise 2A, Page 32
Question 4
Exercise 2B, Page 35
Questions 1,2,5
Worked Examples 2.1 to 2.4
Lesson 2
The Sum of Independent Poisson Variables
For independent variables, X and Y, if 𝑋 ̴ 𝑃𝑜 (𝜆1 ) and 𝑌 ̴ 𝑃𝑜(𝜆2 ) then
𝑋 + 𝑌 ̴ 𝑃𝑜 (𝜆1 + 𝜆2 ).
Example:
Two identical racing cars are being tested on a circuit. For each car, the number of
mechanical breakdowns can be modelled by a Poisson distribution with a mean of one
breakdown in 100 laps. If a car breaks down it is attended and continues on the
circuit. The first car is tested for 20 laps and the second car for 40 laps.
Find the probability that the service team is called out to attend to breakdowns
a) once, b) more than twice
Answers:
a) 0.329 b) 0.023
HW/ CW
Worksheet
Q2
End of chapter review exercise 2, Page 47
Q6,7,8,9
Lesson 3
Poisson Approximation to the Binomial
Distribution
If n is large (𝑛 > 50) and p is small (𝑛𝑝 < 5), the binomial distribution 𝑋 ̴ 𝐵 (𝑛, 𝑝) can be
approximated using a Poisson distribution with mean 𝑛𝑝.
Note that the approximation gets better as 𝑛 gets larger and 𝑝 smaller.
Q1
A Christmas draw aims to sell 5000 tickets, 50 of which will win a prize
a) A syndicate buys 200 tickets. Let X represent the number of these tickets that win a prize.
(i) Justify the use of the Poisson approximation for the distribution of X
(ii) Calculate P (X ≤ 3).
b) Calculate how many tickets should be bought in order for there to be a 90% probability of
winning at least one prize.
Answers:
a) (ii) 0.86 b) 231
Q2
(i) 0.777
(ii) 0.398
Lesson 4
The Normal approximation to the
Poisson distribution
If 𝑋 follows a Poisson distribution with parameter 𝜆, i.e. 𝑋 ̴𝑃𝑜 (𝜆), then 𝐸 (𝑋) = 𝜆 and 𝑉𝑎𝑟 (𝑋) = 𝜆
When 𝜆 is large ( say 𝜆 > 15), the normal distribution can be used as an approximation, where
𝑋 ̴ 𝑁(𝜆, 𝜆).
A continuity correction is needed since we are using a continuous distribution as an approximation to a
discrete one.
Q1
A radioactive disintegration gives counts that follow a Poisson distribution with a mean count
of 25 per second.
Find the probability that in a one-second interval the count is between 23 and 27 inclusive.
Answer:
0.383
HW/ CW
Worksheet
4,6,7,8,10
End of chapter review exercise 2, Page 48
Q10, 13
Worked Examples 2.5 and 2.6
Hypothesis Testing with the Poisson
Distribution
How is the random variable X defined for a Poisson distribution?
What are the assumptions for a Poisson distribution?
What is the probability distribution function for a Poisson
distribution?
What is E(X) and Var(X)?
What is the condition for using a Normal distribution as an
approximation to the Poisson distribution?
Question 1
Question 2
Question 3
The number of calls to a consumer hotline can be modelled by a Poisson
distribution with mean 62 calls every 5 minutes. Salina believes this
average is too low and observes the number of calls recorded during a
randomly chosen 5-minute interval to be 70. Stating the null and
alternative hypotheses, test Salina’s belief at the 10% significance level.
Question 4
A small shop sells, on average, seven laptops per week. Following a
price rise, the number of laptops sold drops to four per week. Test at
the 5% significance level whether the sales of laptops have significantly
reduced
Question 5
At a certain company, machine faults occur randomly and at a constant
mean rate of 1.5 per week. Following an overhaul of the machines, the
company boss wishes to determine if the mean rate of machine faults
has fallen. The number of machine faults recorded over 26 weeks is 28.
Use a suitable approximation and test at the 5% significance level
whether the mean rate has fallen.
Question 6
The number of misprints per page of a newspaper follows a Poisson
distribution with mean two per page. Following new procedures, 49
misprints are found in 32 pages of the newspaper.
a. Use a suitable approximation to test at the 5% level of significance if
the mean number of misprints has changed.
b. How many misprints would be needed for a Type I error to have
occurred?
Solution
Practice
End of chapter review exercises 2
Page 48
Questions 11,12
Worked Example 2.7 and 2.8