Premium
Neural network modeling of physical properties of chemical compounds
Author(s) -
Kozioł J.
Publication year - 2001
Publication title -
international journal of quantum chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.484
H-Index - 105
eISSN - 1097-461X
pISSN - 0020-7608
DOI - 10.1002/qua.1313
Subject(s) - quantitative structure–activity relationship , artificial neural network , boiling point , experimental data , biological system , molecular descriptor , set (abstract data type) , chemistry , data set , computer science , artificial intelligence , machine learning , mathematics , organic chemistry , statistics , biology , programming language
Three different models relating structural descriptors to normal boiling points, melting points, and refractive indexes of organic compounds have been developed using artificial neural networks. A newly elaborated set of molecular descriptors was evaluated to determine their utility in quantitative structure–property relationship (QSPR) studies. Applying two data sets containing 190 amines and 393 amides, neural networks were trained to predict physical properties with close to experimental accuracy, using the conjugated gradient algorithm. Obtained results have shown a high predictive ability of learned neural networks models. The fit error for the predicted properties values compared to experimental data is relatively small. © 2001 John Wiley & Sons, Inc. Int J Quant Chem 84: 117–126, 2001