Random Diffusion Representations
Moshe Salhov, Amir Z. Averbuch · 2019
The diffusion maps is a kernel based method for manifold learning and data analysis that models a Markovian process over data. Analysis of this process provides meaningful information about the inner geometric structures in the data.In this paper, we present a representation framework for analyzing datasets. This framework is based on a random approximation of the diffusion maps kernel. The resulted representation approximate the pair-wise diffusion distance, does not depend on the data size and it is invariant to scale.