To create Gaussian quadrature points, use rule="golub_welsch" instead of "C". Or if you know that you are sticking to Uniform inputs, use the dedicated and somewhat faster Gauss-Legendre method directly with rule="gauss_legendre".
It will automatically create tensor-products for multivariate distributions.
If I understand correctly, the answer is yes. It is one of the main usecases of Chaospy.
See here: http://chaospy.readthedocs.io/en/master/tutorial.html#pseudo-spectral-projection-method
To create Gaussian quadrature points, use
rule="golub_welsch"instead of"C". Or if you know that you are sticking to Uniform inputs, use the dedicated and somewhat faster Gauss-Legendre method directly withrule="gauss_legendre". It will automatically create tensor-products for multivariate distributions.Originally posted by @jonathf in https://github.com/jonathf/chaospy/issues/95#issuecomment-398317557