Forecasting Nigeria’s Inflation Components Using Machine Learning and Econometric Models: A Comparative Analysis

This study models and forecasts monthly headline, food, and core inflation in Nigeria using both traditional econometric and modern machine learning approaches. Specifically, the Autoregressive Integrated Moving Average (ARIMA) and Generalized Autoregressive Conditional Heteroskedasticity (GARCH) models were compared with Random Forest (RF) and Extreme Gradient Boosting (XGBoost) algorithms to evaluate their predictive accuracy and stability. Monthly inflation data obtained from the Central Bank of Nigeria and the National Bureau of Statistics were analyzed, covering food, core, and overall inflation components. The results reveal that while ARIMA and GARCH effectively capture linear and volatility dynamics, machine learning models particularly XGBoost exhibited superior predictive performance in out-of-sample forecasts. Among the inflation components, food inflation emerged as the dominant driver of overall price volatility, underscoring its structural and policy significance. The study concludes that integrating machine learning models into inflation forecasting frameworks enhances predictive precision and early-warning capacity for policy authorities. Consequently, policymakers are encouraged to strengthen data analytics infrastructure, stabilize food supply chains, and adopt data-driven monetary and fiscal interventions to mitigate inflationary pressures and support food security in Nigeria.