Sadiq Zubair, Yang Wenhong, Meraz Md Mostakim, Yang Weisheng, Sun Wen-Hua
Key Laboratory of Engineering Plastics, Beijing National Laboratory for Molecular Science, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China.
University of Chinese Academy of Sciences, Beijing 100049, China.
Molecules. 2024 May 15;29(10):2313. doi: 10.3390/molecules29102313.
In anticipation of the correlations between catalyst structures and their properties, the catalytic activities of 2-imino-1,10-phenanthrolyl iron and cobalt metal complexes are quantitatively investigated via linear machine learning (ML) algorithms. Comparatively, the Ridge Regression model has captured more robust predictive performance compared with other linear algorithms, with a correlation coefficient value of 0.952 and a cross-validation value of 0.871. It shows that different algorithms select distinct types of descriptors, depending on the importance of descriptors. Through the interpretation of the model, the catalytic activity is potentially related to the steric effect of substituents and negative charged groups. This study refines descriptor selection for accurate modeling, providing insights into the variation principle of catalytic activity.
鉴于催化剂结构与其性能之间的相关性，通过线性机器学习（ML）算法对2-亚氨基-1,10-菲咯啉铁和钴金属配合物的催化活性进行了定量研究。相比之下，岭回归模型与其他线性算法相比具有更强的预测性能，相关系数值为0.952，交叉验证值为0.871。结果表明，不同的算法根据描述符的重要性选择不同类型的描述符。通过对模型的解释，催化活性可能与取代基的空间效应和带负电荷的基团有关。本研究优化了描述符选择以进行准确建模，为催化活性的变化原理提供了见解。
