What is it about?
We compare modern forecasting tools to predict futures prices for key commodities (oil, gas, precious metals). Using data from a calm year (2018) and a turbulent year (2022), we test decision-tree models versus neural networks over short (21-day) and medium (125-day) horizons, and report which methods stay most accurate.
Featured Image
Why is it important?
Commodity price swings raise costs and earnings volatility, especially in shocks like 2022. This study is timely because it benchmarks “classic” tree-based models against AI neural networks in both calm and volatile environments, across two practical horizons. It shows AI models deliver noticeably lower errors and can support enterprise risk management and hedging decisions by improving purchase-price planning.
Perspectives
What stood out to me is how much the economic environment matters: forecasting is simply harder in volatile times, so robustness is the real test. I also liked the practical framing linking forecast accuracy to hedging and purchasing decisions, not just statistics. The result that GRU slightly beats LSTM is a reminder that small architecture choices can pay off.
Tibor Bareith
ELTE Centre for Economic and Regional Studies Hungary
Read the Original
This page is a summary of: Evaluating the Effectiveness of Modern Forecasting Models in Predicting Commodity Futures Prices in Volatile Economic Times, Risks, January 2023, MDPI AG,
DOI: 10.3390/risks11020027.
You can read the full text:
Contributors
The following have contributed to this page
