I am wondering if this definition of a counting process, implies some properties of the probability distribution associated with the counting process $\{N(t): t \ge 0\}$.


  1. N(0)=0

  2. The process is stationary and and has independent increments.

Does this definition imply:

  1. The only parameter in the probability distribution for $N(t_{1})-N(t_{2})$ is $t_{1}-t_{2}=\Delta t$.

  2. For any set of countable disjoint time increments: $\{[t^{1}_{1},t^{1}_{2}],[t^{2}_{1},t^{2}_{2}],[t^{3}_{1},t^{3}_{2}]........\}$, with lengths $\{\Delta t_{1},\Delta t_{2},\Delta t_{3}......\}$, and $\Delta T = \Sigma \Delta t_{i}$:

$P(N(\Delta T)=n) = \Sigma_{i,j,k....|i+j+k+...=n}[P(N(\Delta t_{1})=i)*P(N(\Delta t_{2})=j)*P(N(\Delta t_{3})=k)*.......]$

Do you agree with this implication?, and if so, is it and equivalence?

  1. is not true. For a Poisson process with parameter $\lambda>0$ the distribution of your difference depends on $\lambda$ as well, so there are more parameters. But there are other processes taking values in $\mathbb{N}$ with stationary and independent increments.

  2. Yes. This is true for processes with stationary and independent increments taking values in $\mathbb{N}$.

  • $\begingroup$ Thank you for your help, I should have been clearer with 1 yes. Maybe one can say that it is only a function of a common parameter, andt $\Delta t$. Do you know if the property (2) has a name? I see that property 2 is sattisfied for a poisson distribution, but for other discrete distributions it may not be sattisfied. $\endgroup$ – user119615 Feb 22 '14 at 10:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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