ch1 Q1.4(ii) it states that when St = X1 + X2 + ... + XNt is a compound Poisson process, if Xj can only take value 0 or 1, then St is also Poisson process. I can understand when Xj = 1 for all j, St is a Poisson Process, because all Xj then sum up to Nt, which is a Poisson Pcoess. But why when some of Xj = 0, St is also Poisson Process? Thanks a lot!
Hi Iamminime Good question! This is similar to the idea of thinning a Poisson process. Say we have that Xj takes the value 0 with probability 50% and 1 with probability 50%. The process St = X1 + X2 + ... + XNt where Nt ~ Pois(lambda * t) is the same as a Poisson process with parameter lambda / 2, ie it is the same as the following process: Rt = Y1 + Y2 + ... + YMt where Yj = 1 with certainty and Mt ~ Pois(lambda / 2 * t) Hope that helps! Andy