Bayesian Learning and Monte-Carlo simulation final project

This repository contains the assignment (with the PDF describing all possible projects all relative datasets), the dataset and main source code of the final project for the Bayesian Learning and Monte-Carlo simulation course.
This project was developed in group (made up of 4) and represented the final examination of the aforementioned course.

The final score obtained with this source code was 30 out of 30 with honours.

Description of the project

The detailed version of the assignment can be found under assignment/Project2025.pdf under paragraph 3.6. It consists in modelling two time-series (US GDP and ICP) in order to make predictions on its future values. In order to do this, a portion of each the time-series was to learn an appropriate modelling, whereas the final part was used to evaluate predictive capabilities of the learned model. Each time-series was modelled in several ways (MA(1), RA(n), ARMA(1, 1), ...) and appropriate (mainly bayesian) metrics were used to evaluate them and perform model selection.
A report describing the whole development process, intuitions and results can be found in this repository.