I have a N-dimensional dataset, for which I need to apply multivariate interpolation. Is there a possible way to use different kind of interpolation methods in different dimensions? I considered to do this since I want to reduce the overall interpolation error, by choosing the interpolation type for each dimension. So far I only found methods which use a single interpolation type (e.g. cubic spline, linear,...) for all dimensions (such as the Matlab function interpn or the Scipy libary, ...). I'm already very thankful for a reference to literature, as I couldn't find anything.