Re: Identifying the distribution of a data set

Its not just the packet sizes, I have other parameters as well like the interarrival times and batch sizes etc. What I am really after is a library routine or a free ware software that gives me the best fit rather than the exact fit.
As you have just mentioned above, after analysis the result is that the distribution is most probably Pareto and for the interarrival times most probably Exponential and so on. So the answers are probable rather than perfect matches.

ali
.