Spectral Graph Embedding and Topological Data Analysis
Mathematics Postgraduate Seminar
11th October 2019, 4:30 pm – 5:30 pm
Fry Building, 2.04
Computing the spectral decomposition of a graph and embedding the resulting eigenvectors into lower dimensions can help us find interesting structure within the network. Even for simple graph models, the resulting point clouds can have interesting topological features where the shape of the data is important, not just the number of clusters.