SpaceFillingMinDist documentation

[mathieu-couplet]

Hello.

Apparently, this class performs a scaling (to get numbers between 0 and 1) along each marginal before computing the minimum (euclidean) distance: this scaling is not documented. This may lead to a wrong use of the class.

The fact that the computed distance is the euclidean one should be indicated.

Thanks,
regards,
Mathieu

[mbaudin47]

@mathieu-couplet : it seems to me that the code uses the normalize method of the sample:

openturns/lib/src/Uncertainty/Algorithm/WeightedExperiments/SpaceFillingMinDist.cxx

Line 62 in a9346d7

const Sample normalizedSample(normalize(sample));

This method does not seem to be available at the Python level, but is implemented here:

openturns/lib/src/Uncertainty/Algorithm/WeightedExperiments/SpaceFillingImplementation.cxx

Line 69 in 1d23922

Sample SpaceFillingImplementation::normalize(const Sample & sample) const

The code is clear: only the min and the max of the sample are used to normalize it (not the distribution as I thought).