moleculeNet

MoleculeNet是一个专门为测试分子性质的机器学习方法而设计的基准测试。为了促进分子机器学习方法的发展，这项工作整理了多个数据集集合，并创建了一套实现许多已知特征化方法和先前提出的算法的软件。所有方法和数据集都集成在开源的DeepChem包中（MIT许可证）。

MoleculeNet is a benchmark specifically designed to evaluate machine learning methods for molecular property tasks. To advance the development of molecular machine learning approaches, this work curated multiple dataset collections and developed a software suite that implements a wide range of established featurization methods and previously proposed algorithms. All methods and datasets are integrated into the open-source DeepChem package, which is released under the MIT License.

搜集汇总

背景与挑战

背景概述

MoleculeNet是一个用于测试分子属性机器学习方法的基准数据集，包含多个标准化数据集和工具，集成在开源的DeepChem包中，旨在加速分子机器学习领域的发展。