Prediction of dielectric dissipation factors of polymers from cyclic dimer structure using multiple linear regression and support vector machine

Abstract

A quantitative structure–property relationship (QSPR) study was performed for the prediction of the dielectric dissipation factors tanδ of polymers by using multiple linear regression (MLR) and support vector machine (SVM) methods. Molecular descriptors of the polymers were calculated from their corresponding cyclic dimer structures. The obtained models with the frequency of measurement (v) and five descriptors as inputs show good predictive ability, and the SVM model predicts the tanδ more accurately than the MLR model. In addition, this QSPR study reflects that the tanδ values of polymers mainly depend on the molecular size and the electronegativities of atoms and substituents as well as the measurement frequency.