Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.



Re: probability density function identification
Posted:
May 25, 2013 5:27 AM


"Ray Koopman" wrote in message news:687ffa75cc0e486ea2587f83227d317b@vy4g2000pbc.googlegroups.com...
On May 24, 2:50 pm, "David Jones" <dajh...@hotmail.co.uk> wrote: > "Ray Koopman" wrote in message > > news:de591757e9da4a24baaa97d3f307c31b@li6g2000pbb.googlegroups.com... > > I took the OP's question to be: using using only two sets of 20,000 > means of 50, one set from each of the two distributions, could we say > which set used which distribution? My conclusion is that we could be > reasonably confident that the set with the algebraically lower skew > used Beta. > =============================================================== > > I think the OP's question was: > using only one set of 20,000 means of 50, decide which of the two > distributions was the original. > > David Jones
OP: "... my question is whether it is possible to examine the vectors RandBeta and RandUnif (without knowing which is which) and determine which was generated from the Beta pdf and which form the Uniform pdf?"
====================================
I was going by the initial statement of the problem by the OP: "I have a question on trying to reverse engineer the probability density function from which a set of numbers were generated", which seemed to imply that there was a single set of results (from a computer program?) and an unknown mechanism by which they had been generated. That seems to me to be the actual dataset to be analysed. Then, separately, there are sets of simulated data based on known distributions ... and these can be used to help the analysis. So the OP may have confused the question to be answered by describing an attempt to answer it.
David Jones



