Forecasting Inflation in Slovakia Using Machine Learning
Year: 2025 Volume: 75 Issue: 4 Pages: 423-438
Abstract: Reliable inflation forecasts are essential for effective policy interventions and maintaining economic stability. This paper evaluates the use of machine learning methods to forecast inflation in Slovakia over different time frames, including the high volatility period of the COVID-19 pandemic and its aftermath. We work with a comprehensive dataset from the National Bank of Slovakia containing 16 years of observations of 380 variables. We find that methods such as Lasso and Boosting, which utilize regularization and address nonlinearities, outperform traditional models in terms predictive power. By deploying nonlinear approaches on a regularized data set, we are able to enhance these positive outcomes further. The implementation of these advanced methods allows policymakers to improve their decision-making capabilities and prepare more accurate inflation forecasts.
JEL classification: E37, E31, C53
Keywords: machine learning, forecasting, inflation, Slovakia, Lasso, Boosting
DOI: https://doi.org/10.32065/CJEF.2025.04.03
 Attachment [PDF]  Data [ZIP] |  Print  Recommend to others |
