Prediction of glass transition temperatures from monomer and repeat unit structure using computational neural networks
Prediction of glass transition temperatures from monomer and repeat unit structure using computational neural networks
复制标题
DOI：
10.1021/6010062o
复制
发表时间：
2002-03-01
期刊：
影响因子：
--
通讯作者：
Jurs, PC
中科院分区：
文献类型：
--
作者：
Mattioni, BE;Jurs, PC
定量结构-性质关系（QSPR）被开发来关联玻璃化转变温度和化学结构。分别使用MSPs和重复单元结构来构建本研究第一部分和第二部分的几个QSPR模型。使用数值描述符开发模型，这些描述符编码有关化学结构（拓扑、电子和几何）的重要信息。在描述符生成之后，使用多元线性回归分析（MLRA）和计算神经网络（CNN）来生成模型。优化例程（模拟退火和遗传算法）被用来找到信息丰富的子集的描述符进行预测。发现10-描述符CNN模型在使用165种聚合物的MPERT结构（第1部分）预测T1值时是最佳的。由10个CNN组成的委员会产生了10.1 K（r（2）= 0.98）的训练集均方根误差和21.7 K（r（2）= 0.92）的预测集均方根误差。使用重复单元结构为251种聚合物开发了11描述符CNN模型（第2部分）。CNN委员会产生了21.1K（r（2）= 0.96）的训练集均方根误差和21.9K（r（2）= 0.96）的预测集均方根误差。
Quantitative structure-property relationships (QSPR) are developed to correlate glass transition temperatures and chemical structure. Both Monomer and repeat unit structures are used to build several QSPR models for Parts I and 2 of this study, respectively. Models are developed using numerical descriptors, which encode important information about chemical structure (topological, electronic, and geometric). Multiple linear regression analysis (MLRA) and computational neural networks (CNNs) are used to generate the models after descriptor generation. Optimization routines (simulated annealing and genetic algorithm) are utilized to find information-rich subsets of descriptors for prediction. A 10-descriptor CNN model was found to be optimal in predicting T, values using the Monomer structure (Part 1) for 165 polymers. A committee of 10 CNNs produced a training set rms error of 10.1 K (r(2) = 0.98) and a prediction set rms error of 21.7K (r(2) = 0.92). An 11-descriptor CNN model was developed for 251 polymers using the repeat unit structure (Part 2). A committee of CNNs produced a training set rms error of 21.1K (r(2) = 0.96) and a prediction set rms error of 21.9K (r(2) = 0.96).