Oil price prediction with NBEATSx using oil news embeddings and S&P 500 prices

Sara Bourbour and MohammadAmin Mohajelin

Abstract

This study investigates oil price prediction by incorporating oil news embeddings and Standard & Poor’s 500 index prices as exogenous variables in neural basis expansion analysis with exogenous variables (NBEATSx), an advanced forecasting model. The findings imply integrating these two exogenous variables into NBEATSx can significantly enhance prediction performance compared with standard approaches. The successful application of NBEATSx using oil news embeddings and Standard & Poor’s 500 index prices constitutes a huge step forward in understanding complex economic events, with implications for financial markets and energy policies. This study contributes to the discourse on oil price forecasting, highlighting the potential of advanced techniques to predict dynamic economic trends.