Release of our models for the prediction of rheological properties of gels from SMILES strings of their parent low molecular weight gelators (LMWG). The models were built using the Bayesian Additive Regression Trees (BART) algorithm. Updated to include 1,3:2,4-Dibenzylidene-d-sorbitol (DBS) LMWG.
This repository contains:
- Dataset used to build the predictive models
- Scripts to train the model locally
- Example notebook for the local prediction of rheological properties from LMWG SMILES
- Google colab implementation of our predictive models for the prediction of rheological properties