The role of news-based sentiment in forecasting crude oil price during the Covid-19 pandemic

During the Covid-19 pandemic, news-based sentiment emerged as a factor linked to crude oil prices in the literature. However, the question remained as to whether this sentiment could be used to more accurately predict crude oil prices. To assess the effect of news-based sentiment on forecasting crude oil prices, five models based on state-of-the-art machine learning methods were compared; they were taken from the literature on crude oil forecasting. Results are reported for each method for the period of the Covid-19 pandemic and also for the years from 1990 to the beginning of the pandemic. This allowed for the examination of the role of news-based sentiment during different periods of economic development and crisis. Across the machine learning methods, a significant effect of news-based sentiment was observed in terms of its predictive performance during the Covid-19 period, in contrast to previous periods, including the financial crisis of 2008-2009.