Michael Baudin

@josephmure Yes, this bug is still alive. No, the bug is not that the file is actually in "Downloads" while I wrote "Téléchargements". The bug is generated by the use...

I updated the script to OT 1.21. Here is the result. The first part is OK. ```python import openturns as ot # Part 1: OK class FUNC(ot.OpenTURNSPythonFunction): def __init__(self): ot.OpenTURNSPythonFunction.__init__(self,...

In order to implement this, I would rather create it as follows: - Create a generic delta method. - Implement the asymptotic distribution of the SRC indices would use the...

The root cause is [Table-maker's_dilemma](en.wikipedia.org/wiki/Rounding#Table-maker.27s_dilemma), i.e. a "5" digit, followed by a long list of zeros. Hence, the result depends on the rounding of the last digit. This failure reveals...

@MerlinKeller : Please review this PR. I tried to take into account for some of your comments.

> I disagree with this change. There _is_ a prior even in the case of LinearLeastSquaresCalibration and NonLinearLeastSquaresCalibration, which is a flat prior emulated using a Normal distribution with a...

This is a reply to @MerlinKeller: Thank you for the review. > Too bad sigma is fixed in the Bayesian calibration case, estimating sigma is cheap & easy and should...

@jschueller : The commit history is a mess at the present time. I can clean that up when requested, but I do not change the history until the review process...

This PR should not be merged, because there is a confusion on the distribution computed by the Gaussian distributions, which sometimes is not the posterior distribution, but is the distribution...

The branch is not lost (I will not delete it), but I have to work before we can submit a new PR. I need to get the covariance matrix of...