Key Laboratory of Engineering Plastics and Beijing National Laboratory for Molecular Science, Institute of Chemistry, Chinese Academy of Sciences, Beijing, China.
University of Chinese Academy of Sciences, Beijing, China.
J Comput Chem. 2020 Apr 30;41(11):1064-1067. doi: 10.1002/jcc.26160. Epub 2020 Feb 5.
This work demonstrates the potential of machine learning (ML) method to predict catalytic activity of transition metal complex precatalyst toward ethylene polymerization. For this purpose, 294 complexes and 15 molecular descriptors were selected to build the artificial neural network (ANN) model. The catalytic activity can be well predicted by the obtained ANN model, which was further validated by external complexes. Boruta algorithm was employed to explicitly decipher the importance of descriptors, illustrating the conjugated bond structure, and bulky substitutions are favorable for catalytic activity. The present work indicates that ML could give useful guidance for the new design of homogenous polyolefin catalyst.
这项工作展示了机器学习（ML）方法在预测过渡金属配合物前催化剂对乙烯聚合催化活性方面的潜力。为此，选择了 294 种配合物和 15 种分子描述符来构建人工神经网络（ANN）模型。所得 ANN 模型可以很好地预测催化活性，并用外部配合物进行了进一步验证。Boruta 算法被用来明确地揭示描述符的重要性，表明共轭键结构和大取代基有利于催化活性。本工作表明，机器学习可以为均相聚烯烃催化剂的新设计提供有用的指导。
