Machine Learning Applications in Inflation Forecasting: A Comparative Study of FAVAR and BVAR Models

Accurate inflation forecasting is critical for policymakers, investors, and financial institutions. Traditional econometric models such as Bayesian Vector Autoregression (BVAR) and Factor-Augmented Vector Autoregression (FAVAR) have been widely used for macroeconomic forecasting. This paper explores the integration of machine learning techniques with these models to enhance prediction accuracy. Using real-world inflation data, we compare the performance of FAVAR and BVAR models with and without machine learning augmentation. Experimental results indicate that hybrid models incorporating machine learning algorithms such as LSTM and XGBoost outperform classical statistical models in terms of predictive accuracy.

Hali tarjima qilinmagan