Régis LEBRUN

Hi Michael, I just read the documentation of the Nataf transformation. I am a little bit surprised by the following points: * what you call the Nataf transformation is only...

Here are some elements: + epsilon can be zero if rho_ is zero + epsilon can be NaN if rho_ contains NaNs The test allows to deal with both cases...

Nice proposal. More generally, we should think of the set of functions we want to have in the C++ library and the set of functions we have in the Python...

Hi Mathieu, Currently, there is no other way to do the job. We could imagine an extension of the KernelMixture class to take into account weights if you just want...

The dimension of a Distribution is a read-only attribute. Don't expect to be able to change it once the distribution has been created. On the other hand, the user should...

So a specific accessor *may* be added to the python distribution object, but in a way such that the dimension remains read-only for all the other distributions. No idea on...

The actual problems are, after a quick inspection: + The BayesDistribution creates a copy of the conditioned distribution for each call of computePDF() + The MarginalDistribution uses the default implementation...

Hi all, The theory of the KS statistics is *very* different depending on the fact that the parameters are known or unknown. - If the parameters are known, then the...

Once again, the main problem is *not* the exact or approximate implementation of Lilliefors's test but the small sample size we want to check: if you take a huge sampling...

@josephmure a phone call would have been more effective regarding this report. IMO there is no need to reopen anything, but to have a meeting on this specific point.