The artificial neural network method is utilized to study the correlation between the overall hydrogen- bond basicity ( ) values of 231 diverse compounds and their quantum chemical parameters. There are nonlinear rela...The artificial neural network method is utilized to study the correlation between the overall hydrogen- bond basicity ( ) values of 231 diverse compounds and their quantum chemical parameters. There are nonlinear relationships of the values with the quantum chemical parameters. The BP neural networks can predict the values when the neural units of the input layers are the quantum chemical parameters. The structures of the BP neural networks or the quantum chemical parameters are different when the compounds belong to different classes.展开更多
The two-dimensional Quantitative Structure-Activity Relationship （2D-QSAR） models have been developed to estimate and predict the inhibitory activities of a series of HEPT analogues against HIV-1 by using quantum ch...The two-dimensional Quantitative Structure-Activity Relationship （2D-QSAR） models have been developed to estimate and predict the inhibitory activities of a series of HEPT analogues against HIV-1 by using quantum chemical parameters and physicochemical parameters. The best model of three parameters yields r = 0.908, r^2A = 0.800 and s = 0.467 based on stepwise multiple regression （SMR） method. The stability of the model has been verified by t-test, and the results show that the model has perfect robustness. The predictive power of QSAR models has been tested by Leave-One-Out （LOO） and Leave-Group（regularly random set）-Out（LGO） procedure Cross-Validation methodology. The r^2cv of 0.755 and r^2pred of 0.759 were obtained, respectively.展开更多
文摘The artificial neural network method is utilized to study the correlation between the overall hydrogen- bond basicity ( ) values of 231 diverse compounds and their quantum chemical parameters. There are nonlinear relationships of the values with the quantum chemical parameters. The BP neural networks can predict the values when the neural units of the input layers are the quantum chemical parameters. The structures of the BP neural networks or the quantum chemical parameters are different when the compounds belong to different classes.
文摘The two-dimensional Quantitative Structure-Activity Relationship （2D-QSAR） models have been developed to estimate and predict the inhibitory activities of a series of HEPT analogues against HIV-1 by using quantum chemical parameters and physicochemical parameters. The best model of three parameters yields r = 0.908, r^2A = 0.800 and s = 0.467 based on stepwise multiple regression （SMR） method. The stability of the model has been verified by t-test, and the results show that the model has perfect robustness. The predictive power of QSAR models has been tested by Leave-One-Out （LOO） and Leave-Group（regularly random set）-Out（LGO） procedure Cross-Validation methodology. The r^2cv of 0.755 and r^2pred of 0.759 were obtained, respectively.
