MULTIPLE LINEAR-REGRESSION (MLR) AND NEURAL-NETWORK (NN) CALCULATIONSOF SOME DISAZO DYE ADSORPTION ON CELLULOSE

Multiple Linear Regression (MLR) analysis and Neural Network (NN) calc ulations are applied to a series of 21 disazo anionic dyes. Three-dime nsional QSAR parameters were derived from the Cartesian coordinates of the dye molecules. Low energy conformations were obtained by molecula r mechanics and quantum chemical calculations. Electronic and steric e ffects in the dye-cellulose binding are present. The proposed MLR mode ls are rough approximations of nonlinear models. Good correlation with the dye affinity from the MLR calculations and a significantly improv ed fitting of the NN over the MLR models are observed. The model valid ity was checked for two proposed models derived from different sets of structural parameters by the leave-one-out cross-validation procedure . For the first model, a better validity ('cross-validated r(2)' value , of 0.622) of the NN model is noticed by leaving out one compound (fo und as outlier) from the training set, in comparison to that of the ML R model obtained for the same set of compounds (q(2) = 0.434). The q(2 ) value of a second MLR proposed model is better than that of the corr esponding NN model. (C) 1997 Elsevier Science Ltd.