The IMA Journal of Management Mathematics, a major operations research journal published by Oxford University Press, awards the best paper prize each year selected by the editors: https://academic.oup.com/imaman/pages/best_paper_prize_winners
In 2020, this prize was awarded to the paper "Forecasting crude oil futures prices using global macroeconomic news sentiment", a joint work of Dr Paresh Date from the Department of Mathematics at Brunel University London, Dr Gautam Mitra from OptiRisk Systems, and former Brunel doctoral researcher Zryan Sadik.
In this work, Dr Date and his co-authors developed a new modelling approach that's capable of reflecting information from macroeconomic news in the forecasting of prices of financial derivatives called futures, which are dependent on crude oil prices.
Futures contracts are far more liquid (i.e. are traded more frequently and in greater volume) than crude oil itself, and they contain more information about market expectations about future crude oil price movements than past data on crude oil prices. An accurate model for forecasting crude oil futures prices enables prudent financial risk management for various oil market participants and enables pricing of 'over the counter' derivative contracts not traded at a regulated exchange, in a consistent fashion with the traded futures contracts.
The novel approach proposed in the paper combines the Kalman filtering method which is popular in econometrics with a parametrised non-linear functional of the volatility factor, which incorporates the macroeconomic news sentiment. This is one of the first attempts at 'explainable AI' in the academic literature on financial mathematics: a traditional time series model driven by clear economic intuition is successfully combined with a non-parametric (or semi-parametric) map to explain macroeconomic news impact, which would otherwise fall under AI domain.
Through a comprehensive numerical experiment, a significant improvement in price prediction is demonstrated through the blending of macroeconomic news and historical market price data.