PMechDB
收藏资源简介:
PMechDB是一个由加州大学欧文分校计算机科学与化学系创建的化学数据集,包含约13000个经过有机化学家团队手动验证的极性基本反应步骤。这些反应步骤不仅平衡且部分原子映射,旨在为机器学习模型提供具有化学严谨性的训练基础。数据集通过收集化学文献和教科书中的反应条目并进行手动编纂而构建,用于预测极性反应机制,推动合成化学领域的创新。
PMechDB is a chemical dataset developed by the Departments of Computer Science and Chemistry at the University of California, Irvine. It contains approximately 13,000 elementary polar reaction steps manually validated by a team of organic chemists. These reaction steps are both balanced and partially atom-mapped, serving as a chemically rigorous training foundation for machine learning models. The dataset was constructed by collecting reaction entries from chemical literature and textbooks and performing manual curation, with the purpose of predicting polar reaction mechanisms and promoting innovation in the field of synthetic chemistry.

- 1Interpretable Deep Learning for Polar Mechanistic Reaction Prediction加州大学欧文分校 · 2025年



