Latent factor in Large-Dymensional datasets: forecasting and data analysis using factor models. An application to the insurance sector

The purpose of this thesis is to retrace the main steps that were taken in the evolution of the factor models and, in addition, to introduce two examples of how to apply the newest techniques developed in such fields to two different typologies of dataset, one traditional, meaning that it is composed mainly by macroeconomic and financial time series, and the other one 'new' which includes time series relevant to the Italian insurance sector and a set of macroeconomic and financial series related to them.