Inflation forecasts using predictive models.

Msc in Data Science.

Outcome:

Graduated 2022, special mention

Summary

Different predictive models are applied to a collection of several time-series of national and international macroeconomic variables. The basis of this research results from the desire to push further the work of almost 10 years in inflation analysis performed at the Central Bank.

SHAP values for the best predictive model found in Luz' work.

Contributions

In order to pursue clarity on the predictive capabilities of the trained models Luz’ work include post-fitting analysis on feature contributions by means of SHAP values .

Tags

Machine Learning, forecasting, ensemble models.