Poisson Random Variables: Scott She Eld
Poisson Random Variables: Scott She Eld
Poisson Random Variables: Scott She Eld
440: Lecture 12
Poisson random variables
Scott Sheffield
MIT
18.440 Lecture 12
1
Outline
18.440 Lecture 12
2
Outline
18.440 Lecture 12
3
Poisson random variables: motivating questions
�
I The number e is defined by e = limn→∞ (1 + 1/n)n .
�
I It’s the amount of money that one dollar grows to over a year
compounded.
�
I Similarly, e λ = limn→∞ (1 + λ/n)n .
�
I It’s the amount of money that one dollar grows to over a year
compounded.
�
I It’s also the amount of money that one dollar grows to over λ
continuously compounded.
�
I Can also change sign: e −λ = limn→∞ (1 − λ/n)n .
18.440 Lecture 12
5
Bernoulli random variable with n large and np = λ
�
I Let λ be some moderate-sized number. Say λ = 2 or λ = 3.
�
I Suppose I have a coin that comes up heads with probability
�
I How many heads do I expect to see?
�
I Answer: np = λ.
�
I Let k be some moderate sized number (say k = 4). What is
�
I Binomial formula:
Bn n k n(n−1)(n−2)...(n−k+1) k
k p (1 − p)n−k = k! p (1 − p)n−k .
k k
I
� This is approximately λk! (1 − p)n−k ≈ λk! e −λ .
�
I A Poisson random variable X with parameter λ satisfies
k
P{X = k} = λk! e −λ for integer k ≥ 0.
18.440 Lecture 12
6
Outline
18.440 Lecture 12
7
Outline
18.440 Lecture 12
8
Probabilities sum to one
�
I A Poisson random variable X with parameter λ satisfies
k
p(k) = P{X = k} = λk! e −λ for integer k ≥ 0.
∞
I
� How can we show that k=0 p(k) = 1?
λk
�
I Use Taylor expansion eλ = ∞ k=0 k! .
18.440 Lecture 12
9
Expectation
�
I A Poisson random variable X with parameter λ satisfies
k
P{X = k} = λk! e −λ for integer k ≥ 0.
�
I What is E [X ]?
�
I We think of a Poisson random variable as being (roughly) a
p = λ/n.
�
I This would suggest E [X ] = λ. Can we show this directly from
the formula for P{X = k}?
�
I By definition of expectation
∞ ∞ ∞
� � λk � λk
E [X ] = P{X = k}k = k e −λ = e −λ .
k! (k − 1)!
λj −λ
Setting j = k − 1, this is λ
∞
P
j! e
= λ.
�
I
j=0
18.440 Lecture 12
10
Variance
k
�
I Given P{X = k} = λk! e −λ for integer k ≥ 0, what is Var[X ]?
�
I Think of X as (roughly) a Bernoulli (n, p) random variable
with n very large and p = λ/n.
�
I This suggests Var[X ] ≈ npq ≈ λ (since np ≈ λ and
q = 1 − p ≈ 1). Can we show directly that Var[X ] = λ?
�
I Compute
∞ ∞ k ∞
2
X
�
2
�
X
2λ −λ
�
X λk−1 −λ
E [X ] = P{X = k}k = k e =λ k e .
k! (k − 1)!
k=0 k=0 k=1
�
I Setting j = k − 1, this is
⎛ ⎞
∞ j
�
X λ
λ ⎝ (j + 1) e −λ
⎠ = λE [X + 1] = λ(λ + 1).
j!
j=0
�
I Then Var[X ] = E [X 2 ] − E [X ]2 = λ(λ + 1) − λ2 = λ.
18.440 Lecture 12 11
Outline
18.440 Lecture 12
12
Outline
18.440 Lecture 12
13
Poisson random variable problems
�
I A country has an average of 2 plane crashes per year.
�
I How reasonable is it to assume the number of crashes is
�
I Assuming this, what is the probability of exactly 2 crashes?
�
I A city has an average of five major earthquakes a century.
Poisson)?
�
I If both candidates average one major gaffe per debate, what is
the probably that the first has at least one major gaffe and
�
I A casino deals one million five-card poker hands per year.
18.440 Lecture 12
14
MIT OpenCourseWare
http://ocw.mit.edu
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.