Comparing Two Empirical CDFs

Hi,

I have a problem where I want to compare an empirical CDF generated from simulation to another empirical CDF that comes from sampled data.

The shape of my simulated CDF changes based on a couple different parameters. I want to then optimize on these parameters in order to get the best match between the two empirical CDFs.

I'm a bit stumped on how to go about doing this. Searching around shows that the KS-Test is a good method for comparing empirical CDFs but the Matlab documentation seems to say that it can only be compared to a PD Object.

Any help would be great.

Thanks

Answer 1

[Tom Lane]

The kstest2 function can work with two samples, using their two empirical CDFs.