**JOINT SEMINAR BETWEEN ICS and IDIDS **

### Approximate Bayesian Computation for large-scale network models

**Antonietta Mira**

Co-director IDIDS, InterDisciplinary Institute of Data Science

Professor of Statistics, USI Universitą della Svizzera italiana

**6 July 2016****14:30****Room 402 Main Building**

Many systems of scientific and societal interest can be investigated as networks. Increasing availability of empirical large-scale data and steady improvements in computational capacity continue to fuel the growth of this field. An aspect that has received less attention is the divide between the two prominent paradigms to the modeling of network structure: the physics based mechanistic approach and the statistical approach. The mechanistic approach typically assumes that microscopic mechanisms governing network formation and evolution are known, and questions often focus on understanding macroscopic features that emerge from repeated application of these known mechanisms. The statistical approach, in contrast, often starts from observed network structures and attempts to infer some aspects about the underlying data generating process.

This talk will focus on mechanistic network models for which there is no closed form expression for the likelihood but thanks to the possibility of easily sampling a network configuration given a set of parameter values, we develop a principled framework, based on Approximate Bayesian Computation, that can be used to bring some of the mechanistic network models into the realm of statistical inference.

An introduction on the very versatile Approximate Bayesian Computation methodology will be given.