Calculation on enantiomeric excess of catalytic asymmetric reactions of diethylzinc addition to aldehydes with topological indices and artificial neural network

The relationships between the enantiomeric excess of products and the structures of the catalysts or reagents in several types of asymmetric diethylzinc addition reactions were studied by using the BP neural network with topological indices. Moderate cross-validated correlation coefficients were yielded by the models generated from the neural network. Then the models were further used to validate reaction mechanisms, determine the proper reaction environment, and predict the activities of new catalysts.