Clustering Analysis on Locally Asymptotically Self-similar Processes
share
In this paper, we design algorithms for clustering locally asymptotically self-similar stochastic processes. We show a sufficient condition on the dissimilarity measure that leads to the consistency of the algorithms for clustering offline and online data settings, respectively. As an example of application, clustering synthetic data sampled from multifractional Brownian motions is provided.
READ FULL TEXT