J. Polanski et al., THE COMPARISON OF MOLECULAR-SURFACES BY NEURAL NETWORKS AND ITS APPLICATIONS TO QUANTITATIVE STRUCTURE-ACTIVITY STUDIES, Quantitative structure-activity relationships, 17(1), 1998, pp. 27-36
Self-organizing Kohonen neural networks can be used to obtain quantita
tive structure-activity relationships. Simple molecules have been sele
cted to discuss the essential features of the transformation. Descript
ors calculated from maps obtained from comparative studies with templa
te molecules applied to the analysis of compounds derived from butane,
as well as ethylcarboxylates and nitro-and cyanoanilines (sweet-tasti
ng compounds) can be correlated with various molecular effects.