Likelihood free methods for inference on complex models

Abstract

Complex models often have intractable likelihoods, so methods that involve evaluation of the likelihood function are infeasible. The aims of the research are • to provide a review of the likelihood free methods (e.g., ABC or synthetic likelihood) used in fitting complex models large dataset; • to use likelihood free methods to make inference on complex models such as random networks models; • to develop the code for the analysis; • to apply the model and methods for networks data from economics and finance such as trade, financial flows networks, financial contagion networks; • to write a final report where methods and results are presented and discussed.