Predicting Gas Prices: How AI Can Help

The Russia-Ukraine conflict has shown just how fragile the energy market can be. When supply chains get disrupted, prices can go wild. This is where AI comes in. It can help predict gas prices, giving governments and traders a heads-up. Most studies on AI and energy prices focus on short-term predictions or compare similar models. But what about predicting prices over a few days? That's where this research steps in. It looks at daily natural gas prices from the Henry Hub in the U. S. , from 1997 to 2024. The goal? To see how well different AI models can predict prices over 1 to 4 days. Four models were put to the test: feedforward neural network, support vector machine, random forest, and long short-term memory (LSTM) network. The results? LSTM came out on top. For one-day predictions, its error rate was just 8. 53%. That's pretty good! So, why does this matter? Well, for governments dealing with gas shortages, LSTM models can help predict price changes a few days in advance. This can help them plan better and avoid supply disruptions. Traders can also use this info to make smarter decisions. In short, this research gives a solid foundation for governments to make energy policies and for investors to trade wisely.