A novel quantitative structure-property relationship （QSPR） model for estimating the solution surface tension of 92 organic compounds at 20℃ was developed based on newly introduced atom-type topological indices. The data set contained non-polar and polar liquids, and saturated and unsaturated compounds. The regression analysis shows that excellent result is obtained with multiple linear regression. The predictive power of the proposed model was discussed using the leave-one-out （LOO） cross-validated （CV） method. The correlation coefficient （R） and the leave-one-out cross-validation correlation coefficient （Rcv） of multiple linear regression model are 0.991 4 and 0.991 3, respectively. The new model gives the average absolute relative deviation of 1.81% for 92 substances. The result demonstrates that novel topological indices based on the equilibrium electro-negativity of atom and the relative bond length are useful model parameters for QSPR analysis of compounds.
Flash point is a primary property used to determine the fire and explosion hazards of a liquid. New group contribution-based models were presented for estimation of the flash point of alkanes by the use of multiple linear regression(MLR)and artificial neural network(ANN). This simple linear model shows a low average relative deviation(AARD) of 2.8% for a data set including 50(40 for training set and 10 for validation set) flash points. Furthermore, the predictive ability of the model was evaluated using LOO cross validation. The results demonstrate ANN model is clearly superior both in fitness and in prediction performance.ANN model has only the average absolute deviation of 2.9 K and the average relative deviation of 0.72%.
A quantitative structure-spectrum relationship (QSSR) model was developed to simulate 13C nuclear magnetic resonance (NMR) spectra of carbinol carbon atoms for 55 alcohols. The proposed model,using multiple linear regression,contained four descriptors solely extracted from the molecular structure of compounds. The statistical results of the final model show that R2= 0.982 4 and S=0.869 8 (where R is the correlation coefficient and S is the standard deviation). To test its predictive ability,the model was further used to predict the 13C NMR spectra of the carbinol carbon atoms of other nine compounds which were not included in the developed model. The average relative errors are 0.94% and 1.70%,respectively,for the training set and the predictive set. The model is statistically significant and shows good stability for data variation as tested by the leave-one-out (LOO) cross-validation. The comparison with other approaches also reveals good performance of this method.
