A New Neural Network Group Contribution Method for Estimation of Upper Flash Point of Pure Chemicals

In this study, a new group contribution-based model is presented for the prediction of the upper flash point temperature of pure compounds based on a large data set containing 1294 pure compounds. The model is a neural network using a number of occurrences of 122 chemical groups in a pure compound to predict its related UFLT (Upper Flash Point Limit). The squared correlation coefficient, average percent error, mean average error, and root-mean-square error of the model over the main data set containing 1294 pure compounds are 0.99, 1.7%, 6, and 8.5, respectively.