Gas Price Prediction Based on Machine Learning Combined with Ethereum Mempool

• Main Authors: Lan, Dongwan, Wang, Hao, Yin, Changchun, Zhou, Lu, Ge, Chunpeng, Lu, Xiaozhen
• Format: Conference Proceeding
• Language: English
• Subjects: Data models; Ethereum; Gas Price; Learning systems; Machine learning; Prediction; Predictive models; Pricing; Proposals; Security
• ISSN: 2155-6814
• DOI: 10.1109/MASS56207.2022.00057

Gas is the internal pricing (metering system) for running a contract or in general any transaction in Ethereum. With the popularity of Ethereum, the deficiency of current Ethereum transaction pricing mechanism First Price Auctions is being amplified. The fee paid to miners is the gas used multiplied by the gas price. Hence, designing an effective and accurate gas price prediction method is of great significance for improving the efficiency, transparency and security of the Ethereum transaction mechanism. After the Ethereum "London" Hard Fork update, EIP-1559 has been proposed to change the historical gas mechanism and make transaction fees less volatile and more predictable. Therefore, we propose a machine learning based method to predict the gas price of next blocks combined with a dynamic feature exploited from mempool after the proposal of EIP-1559. Specifically, we consider the pending transactions and their gas price in the mempool and take them as a machine learning feature for the first time. Due to the update brought by EIP-1559, we refine more features than the related works. We use machine learning models combined with the mempool features for prediction. Experiments conducted on the dataset manifest that our model combined with the mempool data shows good prediction performance, especially significantly improving the two indicators MAE and RMSE. Furthermore, we analyze and discuss the challenges of our scheme and the potential profound effects brought by our work.