摘要
采用径向基函数网络模型,对36种取代苯胺和苯酚类化合物的急性毒性进行预测,网络自相容能力和交叉检验结果良好.结果表明:①采用分子连接性指数、线性溶解能关系参数、正辛醇/水分配系数作为输入特征参数的神经网络预测结果优于仅采用分子连接性指数作为输入特征参数的神经网络预测结果;②采用分子电性距离矢量和正辛醇/水分配系数作为输入特征参数的神经网络预测结果优于仅采用分子电性距离矢量输入特征参数的神经网络预测结果.该方法还可望成为对有机化合物其它性质进行预测的一种有效手段.
The RBF neural networks were used to predict the toxicity of 36 substituted anilines and phenols.The prediction was examined by a self-consistency test and a cross-validation test.The self-consistency test and the cross-validation test gave good results.The calculated results showed that：① the prediction of the neural network obtained by using molecular connectivity indices,linear salvation energy relationships parameters and octanol/water partitioning coefficients as net inputs was better than that of the neural network obtained by using molecular connectivity indices as the net inputs;② the prediction of the neural network obtained by using molecular electronegativity-distance vector and octanol/water partitioning coefficients as the net inputs was better than that of the neural network obtained by using molecular connectivity indices as the net inputs.The ANN method could be used to predict other important properties of organic compounds.
出处
《云南大学学报（自然科学版）》
CAS
CSCD
北大核心
2010年第6期685-689,694,共6页
Journal of Yunnan University(Natural Sciences Edition)
基金
河南省教育厅自然科学研究计划项目资助(2009B150023)
