FinML-Chain
收藏资源简介:
FinML-Chain是由杜克昆山大学开发的区块链集成数据集,旨在增强金融机器学习。该数据集整合了高频链上数据和低频链下数据,提供了经济机制设计研究的新基准。数据集内容包括以太坊的区块链数据和Discord上的用户讨论文本,总大小为80.4MB和4.92MB、13.4MB。数据集的创建过程结合了区块链技术的透明性、不可篡改性和实时更新特性,确保数据的高质量和可靠性。该数据集主要应用于金融市场的机器学习模型优化,特别是以太坊交易费机制的研究,旨在解决数据缺失、透明度不足等问题,提升预测模型的准确性和可靠性。
FinML-Chain is a blockchain-integrated dataset developed by Duke Kunshan University, designed to advance financial machine learning research. This dataset integrates high-frequency on-chain data and low-frequency off-chain data, serving as a novel benchmark for studies on economic mechanism design. The dataset encompasses Ethereum blockchain data and user discussion texts sourced from Discord, with sizes of 80.4 MB, 4.92 MB, and 13.4 MB respectively. During its development, the characteristics of blockchain including transparency, immutability and real-time update are leveraged to guarantee the high quality and reliability of the dataset. This dataset is primarily applied to optimize machine learning models for financial markets, especially research on Ethereum transaction fee mechanisms, with the goals of addressing issues like data scarcity and insufficient transparency, and enhancing the accuracy and reliability of prediction models.

- 1FinML-Chain: A Blockchain-Integrated Dataset for Enhanced Financial Machine Learning杜克昆山大学 · 2024年



