Add a ShuffleSplit method

[vchabri]

What is the idea?

It would be valuable and useful to have a "ShuffleSplit" method such as the one available in Scikit-learn in order to split a Sample into, e.g., two subsamples based on a random shuffling strategy.

Why is this needed?

Such a task is very useful for metamodel training and validation.

Additional Context

Here are the main references:

• Doc page
• Source code

[mbaudin47]

A workaround is to use the method presented in plot_randomize_sample_lines.html. This essentially amounts to using KPermutationsDistribution.