Michael Baudin

Here is a summary of the issues. When we perform Bayesian calibration, there are two different distributions of the parameters: - the posterior ; - the distribution of the MAP....

After some thinking on this topic, it appears to me that, even with limitations, the PR improves on the current implementation, theory and examples. This is why I reopen the...

There is an ongoing work on this topic by @adutfoy and me.

I agree that an option to configure that would be useful. In the meantime, you can already do it by configuring the `markersize` option: ```python R = ot.CorrelationMatrix(2) R[1, 0]...

@m-balesdent suggests to clarify the available sensitivity analysis features in the API section. Moreover, a guide to sensitivity analysis features would be useful. @JPelamatti : Do you have any suggestion...

I would like to illustrate the facts presented by @mdelozzo . - The main help API page is as follows.  From there, it is not straightforward to find the...

As mentioned in #2292, the current equations in `MetaModelResult` seems wrong to me. Here is a more detailed analysis. - [FunctionalChaosAlgorithm, line 384](https://github.com/openturns/openturns/blob/fe388ec3e7a8e23ba1a3f7eada320165cba86fca/lib/src/Uncertainty/Algorithm/MetaModel/FunctionalChaos/FunctionalChaosAlgorithm.cxx#L384C14-L384C33) computes the `residuals` from the `ProjectionStrategy.getResidual()`. -...

In DistributionImplementation, the `quantileEpsilon_` attribute is read from the `ResourceMap`: https://github.com/openturns/openturns/blob/95f0e0e09ecad661f865c6dfbcc46a3fbe518cb9/lib/src/Uncertainty/Model/DistributionImplementation.cxx#L102 In the `TruncatedDistribution.computePDF()` method, the `quantileEpsilon_` attribute is used: https://github.com/openturns/openturns/blob/95f0e0e09ecad661f865c6dfbcc46a3fbe518cb9/lib/src/Uncertainty/Distribution/TruncatedDistribution.cxx#L318 The code: ```cpp x < getRange().getLowerBound()[0] - quantileEpsilon_ ```...

I think that you did not see a couple of suggestions that are hidden by Github, as you did not take them into account :  Is that correct? Please...

- Notice that the doc may be updated afterwards, when the PR #2650 is merged. - The CT is OK for merging. If @josephmure is not OK with that, please...