ComposiLator
Description
This repository contains all datasets and Python source code used to develop, train, and validate ComposiLator, a hybrid AI–mechanistic framework for experiment-free prediction of elastic, shear, and bulk moduli of polymer composites. It includes literature-derived experimental data, descriptor files, preprocessing scripts, and the full prediction workflow accompanying the associated manuscript. A few shear and bulk moduli in this dataset were calculated from reported Young’s modulus and Poisson’s ratio using standard isotropic elasticity relations.
Files
Steps to reproduce
All experimental data were extracted from published literature, as cited in each dataset file. The calculated values (Voigt, Reuss, VRH midpoint, correction factor, Halpin-Tsai predictions and final hybrid-model predictions) can be fully reproduced by following the instructions and running the scripts provided in the accompanying README file.