Sofia Olhede: Network Data Analysis

Tid: On 2018-02-07 kl 15.15 - 17.00

Föreläsare: Sofia Olhede, University College London

Plats: Oskar Klein (AlbaNova)

Schedule 

15:15-16:15 Colloquium lecture by Sofia Olhede in Room Oskar Klein, AlbaNova

16:15-17:00 SMC social get together with refreshments 

Abstract

A network is a set of entities (vertices) and binary relationships (edges) between them. Statistical analysis of such objects is hindered by the fact that the vertex labelling of the network normally does not carry any information, the network is made from a very large set of vertices, and the number of network edges is relatively modest. A natural assumption to make in this setting is that the network is a realisation of an exchangeable array. I will discuss different methods of data analysis for such types of data with weak assumptions on the data generating mechanism, and what information can be gleaned from different data summaries that both carry structural information, and a notion of scale.