Interest-Rate-Forecasting

This repository is dedicated to interest rate forecasting, utilizing models from mathematical finance (Vasicek and CIR) and Machine Learning (Bagging and LSTM). The primary goal is to assess the forecasting performance relative to a Random Walk benchmark across different time horizons: 1 day, 1 week, 1 month, and 3 months.

Overview

Interest rate forecasting plays a pivotal role in financial analysis. This project delves into the realm of interest rate prediction, exploring and evaluating various models. The models employed include mathematical finance models like Vasicek and CIR, as well as machine learning models such as Bagging and LSTM.

Files

Tesi.ipynb

This Jupyter notebook provides an in-depth analysis of interest rate forecasting. It utilizes the Least Squares Method to calibrate Vasicek and CIR models and presents a detailed examination of the results.

CIR_calibration.py

This file contains the implementation of a Maximum Likelihood Estimation approach for calibrating the Cox-Ingersoll-Ross (CIR) model.

Vasicek_calibration.py

In this file, you'll find the implementation of a Maximum Likelihood Estimation approach for calibrating the Vasicek model.