Hodge-Edge-GP

This implements the Gaussian processes on the edge space of a simplicial 2-complex, including applications in forex prediction, ocean current interpolation and water networks state estimation. Our way of constructing the edge Gaussian processes allows separate modeling of the different parts of the edge signals, e.g., divergence-free part or curl-free part. This code shows how we can use Gaussian process regression to interpolate or predict signals on the edge space while respecting their intrinsic properties.