期刊
EUROPEAN JOURNAL OF MEDICINAL CHEMISTRY
卷 45, 期 11, 页码 4856-4862出版社
ELSEVIER FRANCE-EDITIONS SCIENTIFIQUES MEDICALES ELSEVIER
DOI: 10.1016/j.ejmech.2010.07.056
关键词
Antagonistic activity; Artificial neural network; Multiple linear regression; Quantitative structure-property relationship; Molecular descriptor
Quantitative structure-activity relationship (QSAR) method was used to predict the pIC(50) value of 58 derivatives of 6-methoxy benzamides in this work. The artificial neural network (ANN) and multiple linear regressions (MLR) were used to construct the non-linear and linear QSAR models, respectively. The standard errors in the prediction of pIC(50) for training, internal and external test sets, are; 0.280, 0.446 and 0.382 by MLR model and are; 0.175, 0.326 and 0.296 by ANN model, respectively. Also these models were further examined by cross-validation methods which produce the statistics of Q(2) = 0.8340 and SPRESS = 0.322 for MLR model and Q(2) = 0.8055 and SPRESS = 0.219 for ANN model. (C) 2010 Elsevier Masson SAS. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据