Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

For a Poisson process show, for $s < t$, that $P(N(s)=k|N(t)=n) = \binom{n}{k} (\frac{s}{t})^k (1-\frac{s}{t})^{n-k}$

share|improve this question

2 Answers 2

up vote 4 down vote accepted

The joint probability mass function for $(N(s), N(t))$ for $s<t$ is, using the property of independent increments, i.e. $N(t) \sim N(s) + N(t-s)$: $$ \mathbb{P}(N(s) = k, N(t) = n) = \mathbb{P}(N(s) = k, N(t) - N(s) = n - k) = \\\frac{(\mu s)^k}{k!} \mathrm{e}^{-\mu s} \cdot \frac{(\mu (t-s))^{n-k}}{(n-k)!} \mathrm{e}^{-\mu(t-s)} \cdot [ n-k \geqslant 0, k \geqslant 0] $$ Then, assuming $n \geqslant k \geqslant 0$: $$ \mathbb{P}(N(s) =k | N(t) = n) = \frac{\mathbb{P}(N(s)=k,N(t)=n)}{\mathbb{P}(N(t)=n)} = \\ \frac{\frac{\mu^k s^k}{k!} \cdot \frac{\mu^{n-k} (t-s)^{n-k}}{(n-k)!} \cdot \mathrm{e}^{-\mu t} }{ \frac{\mu^n t^n}{n!} \mathrm{e}^{-\mu t}} = \left( \frac{s}{t} \right)^k \left( 1- \frac{s}{t} \right)^{n-k} \binom{n}{k} $$

share|improve this answer
    
+1. Neat'n easy. –  Did Mar 16 '12 at 0:20

Let

$\ \ \ \ A$ be the number of events occurring in the interval $[0,s]$,

$\ \ \ \ B$ be the number of events occurring in the interval $[s,t]$,

and

$\ \ \ \ C$ be the number of events occurring in the interval $[0,t]$.

Using the fact that in a Poisson process, events occurring in one time frame are independent of events occurring in a different time frame, we may write the sought after probability as $$\tag{1} \eqalign{P ( A =k \mid C= n) &={P\bigl((A=k)\cap ( C=n) \bigr)\over P(C=n)}\cr &={P\bigl((A=k)\cap ( B=n-k) \bigr)\over P(C=n)}\cr &={P (A=k) P ( B=n-k) \over P(C=n)} .} $$

Now, if the Poisson process has parameter $\lambda$, then

$\ \ \ \ A$ has Poisson distribution with parameter $s\lambda$,

$\ \ \ \ B$ has Poisson distribution with parameter $(t-s)\lambda$,

and

$\ \ \ \ C$ has Poisson distribution with parameter $t\lambda$.

So, using equation $(1)$: $$\eqalign{ P( A =k \mid C= n) &={ {(s\lambda)^k e^{-s\lambda}\over k!} {( (t-s)\lambda)^{n-k} e^{-(t-s)\lambda}\over (n-k)!} \over {(t\lambda)^n e^{-t\lambda}\over n!}}\cr &={n\choose k} {s^k(t-s)^{n-k} \lambda^k \lambda^{n-k} e^{-t\lambda}\over t^n\lambda^ne^{-t\lambda}}\cr &={n\choose k} {s^k(t-s)^{n-k}\over t^kt^{n-k}}\cr &={n\choose k} {\Bigl({s\over t}\Bigr)^k \Bigl(1-{s\over t}\Bigr)^{n-k}}.\cr } $$

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.