Barycentric Transformation#
In one dimension, barycentric Lagrange interpolation is the most efficient
scheme for fixed interpolation nodes[1].
Both determining the interpolating polynomial
Minterpy has partially extended the classic barycentric Lagrange interpolation
to the multidimensional case.
This extension leverages the fact that the transformation
from the Lagrange to Newton basis
involves structured sparse triangular matrices.
Exploiting this structure, as indicated by our preliminary results
for the case of
In summary, we aim to reduce the runtime complexity from
References