Application of machine learning approach for prediction the heat capacity of amine

This study integrates computational chemistry and machine learning to investigate the relationship between the chemical functional groups and the heat capacity values of amine compounds. The heat capacity of amine compounds was estimated using a hybrid method that includes a simple group contribution (GC) method implemented in a radial basis function neural network (RBF NN). Genetic function approximation (GFA) as a proper computational method was used for selection the most important functional groups and linear model developing. The nonlinear relation between the selected functional groups and the heat capacity values of amine compounds was determined by RBF NN. The validation of GC models illustrated that the squared correlation coefficient (R۲) between predicted and experimental values were ۰.۹۲۹ and ۰.۹۵۴ for GFA and RBF NN, respectively. The obtained results in this article suggest that by using machine learning approach, it is possible to obtain a good estimation of the liquid heat capacity values of amine compounds.

کلیدواژه ها:

Amine ، Genetic function approximation (GFA) ، Group contribution (GC) ، Heat capacity ، Machine learning ، Radial basis function neural network (RBF NN)