How to select oil price prediction models — The effect of statistical and financial performance metrics and sentiment scores

Abstract

Predicting crude oil prices is an important yet challenging forecasting problem due to various influencing quantitative and qualitative factors. To address the growing number of potential prediction models and model parameters to consider during model selection, we highlight the need to systematically compare alternative prediction models and variables while taking the specific context of their application into account. Specifically, we provide a novel perspective on oil price prediction models by comparing a variety of different forecasting models and considering both their statistical and financial performance. In addition to common statistical measures, to assess the usefulness in a practical setting we evaluate the potential financial impact of the predictions in a simulation of a simple trading strategy. We show that the ranking of different approaches depends on the selected evaluation metric and that small differences between models in one evaluation metric can translate into large differences in another metric. For instance, forecasts that are not considered statistically different can lead to substantially different financial performance when the forecasts are used in a trading strategy. Finally, we show that including qualitative information in the prediction models through sentiment analysis can yield both statistical and financial performance improvements.