This project aims at predicting the evolution of ECB's deposit rate 12-month ahead using macroeconomic variables and models such as Linear Regression, Ridge Regression, Random Forest and LightGBM. We also use an ensemble stacking model to combine those models and "diversify" their predictive power based on learned weights. Programmed in R.