Poisson model and probabilities

Poisson model and probabilities

Spec SB1–SB3 · AQA A-Level Further Maths: Revision notes

Studying this course? Keep your progress in one place.

Key facts

  • X∼Po(λ)X\sim\mathrm{Po}(\lambda) models events that occur singly, independently and at a constant average rate.
  • P(X=r)=e−λλrr!\mathrm{P}(X=r)=\dfrac{e^{-\lambda}\lambda^r}{r!}.
  • For "at least", use the complement: P(X≥r)=1−P(X≤r−1)\mathrm{P}(X\ge r)=1-\mathrm{P}(X\le r-1).
  • Mean == variance =λ=\lambda; standard deviation λ\sqrt\lambda.
  • λ\lambda scales with the interval: 2.52.5 per metre gives 55 in 22 m.

When does a Poisson model apply?

When events occur singly, independently and at a constant average rate in a fixed interval.

A Poisson distribution models the number of times an event occurs in a fixed interval of time, length, area or volume. Write X∼Po(λ)X\sim\mathrm{Po}(\lambda), where the parameter λ>0\lambda>0 is the average number of occurrences in the interval.

The model is valid when events occur singly, independently and at a constant average rate.

Examples: faults in fabric, calls to a call centre. XX has no upper limit: 0,1,2,3,…0,1,2,3,\dots

00.050.10.150.20.250.0120.003401234567Number of faults, xP(X = x)
Po(2): the probability of X faults in a fixed length. X has no upper limit, so the probabilities tail off but never reach zero.

Poisson fits

Singly:
One fault at a time
Independently:
Calls unrelated
Constant rate:
Steady average per hour

Poisson fails

Singly:
Buses arrive in pairs
Independently:
One event triggers another
Constant rate:
Rush hour, seasons

Which condition would be broken if buses tend to arrive in bunches?

The Poisson formula

Use P(X=r)=e−λλrr!\mathrm{P}(X=r)=\frac{e^{-\lambda}\lambda^r}{r!}, or the calculator's Poisson functions.

If X∼Po(λ)X\sim\mathrm{Po}(\lambda) then

P(X=r)=e−λλrr!,r=0,1,2,…\mathrm{P}(X=r)=\frac{e^{-\lambda}\lambda^{r}}{r!},\qquad r=0,1,2,\dots

This is the probability mass function. On a calculator use the Poisson probability function for P(X=r)\mathrm{P}(X=r) and the cumulative function for P(X≤r)\mathrm{P}(X\le r). Check your calculator against the formula for one value.

00.050.10.150.20.250.009901234567rP(X = r)
Distribution of X ~ Po(2.5), probabilities to 4 decimal places

Worked example

X∼Po(2.5)X\sim\mathrm{Po}(2.5). Find P(X=2)\mathrm{P}(X=2).

For X∼Po(2.5)X\sim\mathrm{Po}(2.5), what is P(X=0)\mathrm{P}(X=0)?

Cumulative and "at least" probabilities

Use the complement: P(X≥r)=1−P(X≤r−1)\mathrm{P}(X\ge r)=1-\mathrm{P}(X\le r-1).

Because XX has no upper limit, "at least" probabilities use the complement. A cumulative probability is P(X≤r)=∑k=0rP(X=k)\mathrm{P}(X\le r)=\sum_{k=0}^{r}\mathrm{P}(X=k).

P(X≥r)=1−P(X≤r−1).\mathrm{P}(X\ge r)=1-\mathrm{P}(X\le r-1).

Also P(X>r)=P(X≥r+1)\mathrm{P}(X>r)=\mathrm{P}(X\ge r+1) and P(a≤X≤b)=P(X≤b)−P(X≤a−1)\mathrm{P}(a\le X\le b)=\mathrm{P}(X\le b)-\mathrm{P}(X\le a-1).

00.20.40.60.81012345678rP(X ≤ r)
Cumulative probabilities P(X ≤ r) for Po(2). For example P(X ≥ 3) = 1 − P(X ≤ 2) = 1 − 0.6767 = 0.3233.
  • P(X≥r)\mathrm{P}(X\ge r)1−P(X≤r−1)1-\mathrm{P}(X\le r-1)
  • P(a≤X≤b)\mathrm{P}(a\le X\le b)P(X≤b)−P(X≤a−1)\mathrm{P}(X\le b)-\mathrm{P}(X\le a-1)

Worked example

X∼Po(1.8)X\sim\mathrm{Po}(1.8). Find P(X≥3)\mathrm{P}(X\ge3).

Which expression equals P(X>4)\mathrm{P}(X>4)?

Mean, variance and standard deviation

The mean and variance are both λ\lambda, so the standard deviation is λ\sqrt\lambda.

If X∼Po(λ)X\sim\mathrm{Po}(\lambda) then E(X)=λ\mathrm{E}(X)=\lambda, Var(X)=λ\mathrm{Var}(X)=\lambda and the standard deviation is λ\sqrt{\lambda}.

The mean and variance are equal. This is a useful check: if data have a sample mean far from the sample variance, a Poisson model is probably unsuitable.

Y∼Po(6)Y\sim\mathrm{Po}(6) has Var(Y)=6\mathrm{Var}(Y)=6 and standard deviation 6=2.45\sqrt6=2.45.

00.020.040.060.080.10.120.140.160.00250123456789101112xP(X = x)
Po(6): mean 6, variance 6, standard deviation √6 = 2.45. The distribution is centred on λ.
  • Meanλ\lambda
  • Varianceλ\lambda
  • Standard deviationλ\sqrt\lambda

X∼Po(9)X\sim\mathrm{Po}(9). What is its standard deviation?

Changing the interval

Because the rate is constant, λ\lambda scales with the interval; an unknown λ\lambda can come from a given probability.

Because the constant average rate is the same throughout, λ\lambda scales with the size of the interval. At 2.5 faults per metre, a 2 m length has λ=5\lambda=5 and a 10 cm length has λ=0.25\lambda=0.25.

Sometimes λ\lambda must be found from given information. If P(X=0)=0.2\mathrm{P}(X=0)=0.2 then e−λ=0.2e^{-\lambda}=0.2, so λ=ln⁡5=1.61\lambda=\ln5=1.61 and the standard deviation is λ=1.27\sqrt\lambda=1.27.

"Within one standard deviation of the mean" means λ−λ<X<λ+λ\lambda-\sqrt\lambda<X<\lambda+\sqrt\lambda.

00.20.40.60.80.00670.03370.03630.02430.0020.00010123456789Number of faults, xP(X = x)
  • 2 m: Po(5)
  • 10 cm: Po(0.25)
λ scales with the interval: 2 m at 2.5 faults per metre gives Po(5); 10 cm gives Po(0.25).
  • New intervalλ=rate×length\lambda=\text{rate}\times\text{length}
  • P(X=0)=p\mathrm{P}(X=0)=pλ=−ln⁡p\lambda=-\ln p

Worked example

Faults occur at an average rate of 2.5 per metre. Find the probability of no faults in a 2 m length.

Worked example

X∼Po(λ)X\sim\mathrm{Po}(\lambda) and P(X=0)=0.2\mathrm{P}(X=0)=0.2. Find the values of XX within one standard deviation of the mean.

Faults occur at 2.5 per metre. What is the mean number of faults in a 40 cm length?

Try an exam question

Calls to a call centre occur singly, independently and at a constant average rate of 1.81.8 per minute. Find the probability that at least 33 calls arrive in a given minute.

[4 marks]

That's the notes covered.

Carry on to the next subtopic.

Try a question

Find P(X=2)\mathrm{P}(X=2).

1 mark, marked against the mark scheme

Revision notes