Accurate prediction of theta (lower critical solution temperature) in polymer solutions based on 3D descriptors and artificial neural networks

作者：Xu, Jie^*; Chen, Biao; Liang, Hao

来源：Macromolecular Theory and Simulations, 2008, 17(2-3): 109-120.

DOI：10.1002/mats.200700067

• artificial neural networks
• quantitative structure-property relationships
• solution properties
• stepwise multilinear regression analysis
• three-dimensional descriptors

Ouantitative structure-property relationships were studied between descriptors representing the three-dimensional structures of molecules and theta (LCST, lower critical solution temperature) in polymer solutions with a database of 169 data containing 12 polymers and 67 solvents. Feed-forward artificial neural networks (ANNs) combined with stepwise multilinear regression analysis (MLRA) were used to develop the models. With ANNs, the squared correlation coefficient (R-2) for theta (LCST) of the training set of 112 systems is 0.9625, the standard error of estimation (SEE) is 13.43 K, and the mean relative error (MRE) is 1.99%; in prediction of theta (LCST) using the test set of 57 systems, the MRE is 2.26%. With MLRA, the MREs for the training % (R-2 and test sets are 4.02% (R-2 = 0.8739, SEE = 25.88 K) and 5.05%, respectively.

• 出版日期2008-4-16
• 单位中国科学技术大学; 武汉纺织大学; 惠州学院