Feature Learning applied to the Estimation of Tensile Strength at Break in Polymeric Material Design.

Several feature extraction approaches for QSPR modelling in Cheminformatics are discussed in this paper. In particular, this work is focused on the use of these strategies for predicting mechanical properties, which are relevant for the design of polymeric materials. The methodology analysed in this study employs a feature learning method that uses a quantification process of 2D structural characterization of materials with the autoencoder method.

Novel descriptors from main and side chains of high-molecular-weight polymers applied to prediction of glass transition temperatures.

New descriptors of main and side chains for polymers with high molecular weight are presented in order to predict the glass-transition temperature (T(g)) by means of T(g)/M ratio. They were obtained by molecular modeling for the middle unit in a series of three repeating units (trimer). Taken together with other classic descriptors calculated for the entire trimeric structure, the ones that correlated better with the property were selected by using a variable selection method.