Association constant prediction for the inclusion of alpha-cyclodextrin with benzene derivatives by an artificial neural network

The association constants (K-a) for the inclusion complexation of alpha-cyc lodextrin (alpha-CD) with 72 mono- and 1,4-disubstituted benzenes were pred icted successfully by an artificial neural network (ANN) with molar refract ion (R-m) and hydrophobic constant (pi) as input parameters, which reflect the volume and hydrophobicity of the substituents respectively. The predict ions strongly suggested that the inclusion complexation of alpha-CD with gu est molecules was mainly driven by van der Waals forces and hydrophobic int eractions