Topology-based tools for data classification
Recently, topological data analysis is gaining a more and more relevant role in the extraction of the core information from large and unorganised datasets. In particular, persistent homology provides a powerful tool enabling a stable data classification.
In this talk, Dr Fugacci will investigate the capability of persistent homology in being integrated with kernel methods in a machine learning framework, mainly focusing on the case of multivariate datasets.
Moreover, he will discuss the possibility of developing new topological tools specifically designed for catching and visualising the global structure of complex networks.