Michael Baudin

The theory [doc](https://openturns.github.io/openturns/latest/theory/probabilistic_modeling/process_definitions.html) of the autocorrelation function of a stochastic process is wrong:  What is wrong is that, in the square root, we should have the product of the...

getSampleAtVertex() is not robust

1

It seems to me that the method `getSampleAtVertex(nodeIndex)` does not check its input. Hence, any integer `nodeIndex` can be used, which may produce memory errors. With OT1.19, the following script:...

Fixed calibration labels

1

When we perform calibration, we use two different classes of algorithms: - probabilistic algorithms such as `LinearLeastSquaresCalibration` and `NonLinearLeastSquaresCalibration, - Bayesian algorithms such as `GaussianLinearCalibration` and `GaussianNonLinearCalibration`. The `drawParameterDistributions()` method...

We cannot compute Wilk's rank

2

The current `Wilks` class: https://openturns.github.io/openturns/latest/user_manual/_generated/openturns.Wilks.html can compute the sample size given the quantile and the rank. In some cases, however, we need to compute the rank given the quantile and...

SRC indices do not have confidence bounds

10

When we estimate SRC indices, we cannot get neither confidence bounds nor the distribution. There are several consequences. - When the number of simulations is set, we cannot know the...

The goal of this PR is to create a new calibration example without observed inputs:  This PR fixes #2030. ~~The changelog should be updated before merging.~~

Smolyak experiment

1

This PR creates a new `SmolyakExperiment` class. It allows to create Smolyak quadratures. The simplest example is: ```python # Generate a Smolyak Gauss-Legendre rule in 2 dimensions. experiment1 = ot.GaussProductExperiment(ot.Uniform(0.0,...

The graph produced by DrawCorrelationCoefficients cannot draw SRC indices

2

Consider the following script. ```python import openturns as ot g = ot.SymbolicFunction(["X1","X2"],["X1*X2"]) X1 = ot.Normal(0.,10.) X2 = ot.Uniform(-1.,1.) X = ot.ComposedDistribution([X1,X2]) inputDescription = ["X1","X2"] X.setDescription(inputDescription) n = 10000 sampleX =...

After several requirements from users, I created an example which shows how to calibrate a parametric model which has no observed input. This was done in the PR #1490. The...

There are several classes which constructor is based on a `Collection`. For such a class, there is often a method which sets the collection. The doc of this method should...