Accurate quantitative structure–property relationship analysis for prediction of nematic transition temperatures in thermotropic liquid crystals

Abstract

The quantitative structure–property relationship approach was performed to study the nematic transition temperatures (T N) in thermotropic liquid crystals. The multi-linear regression analysis (MLRA) and artificial neural networks (ANNs) were employed to develop linear and nonlinear models, respectively. The proposed linear model contains five descriptors, with the squared correlation coefficient R 2 of 0.9837 and the standard error of estimation s of 2.31. The mean relative errors (MREs) for the training and test sets are 3.15 and 5.21%, respectively. Better predictive results were obtained from the nonlinear model: the MREs for the training and test sets are 2.03% (R 2 = 0.9911 and s = 1.71) and 1.92% (R 2 = 0.9892), respectively.