Use of a neural network to determine the boiling point of alkanes

• DOI: 10.1039/ft9949000097
• Publication Date: 2004-04-15T13:04:27Z
• Authors: Driss Cherqaoui; Didier Villemin
• Keywords: Boiling point; Network Structure; Training set; Data set; Backpropagation

Back-propagation neural networks (NNs) are useful for the study of quantitative structure–activity relationships or structure–property correlations. Models between structure and boiling point (bp) 150 alkanes were constructed by means a multilayer network (NN) using back-propagation algorithm. The results our NN compared with those other models from literature, found to be better. points then predicted removing 15 compounds (test set) 135 molecules as training set. Using same process, all in data bank groups compounds. obtained satisfying.