人工神经网络用于有机磷酸酯类定量结构活性关系研究

采用人工神经网络(ANN)建立了35种有机磷酸酯类化合物的结构及其对家蝇急性毒性之间的定量关系模型(ANN模型),以35种有机磷酸酯类化合物的分子电性距离矢量作为输入、对家蝇的急性毒性作为输出,采用内外双重验证的办法分析和检验所得模型的稳定性,所构建网络模型的相关系数为0.999 9、交叉检验相关系数为0.995 8、标准偏差为0.114 1、残差绝对值≤0.40,应用于外部预测集,外部预测集相关系数为0.986 0。而多元线性回归(MLR)法模型的相关系数为0.976 0、标准偏差为0.256 8、残差绝对值≤0.57,外部预测集相关系数为0.975 8。结果表明,ANN模型获得了比MLR模型更好的拟合效果。

QSAR Study of Organophosphate Compounds Using Artificial Neural Network

The systematic study of the quantitative structure-activity relationship(QSAR) on 35 organophosphate compounds was performed based on the artificial neural network.For the artificial neural network method,when using the molecular electronegativity-distance vector as the inputs of the neural network and the acute toxicity to housefly as the outputs of the neural network,the correlation coefficient of established model was 0.999 9,the leave one out cross-validation regression coefficient was 0.995 8,the standard error was 0.114 1,the correlation coefficient of the test set was 0.986 0 and the absolute values of residual were less than or equal to 0.40.In order to make contrast,the QSAR model was set up by multiple linear regressions(MLR) method.For the model built by MLR,the correlation coefficient was 0.976 0,the standard error was 0.256 8 and the absolute values of residual were less than or equal to 0.57,the correlation coefficient of the test set was 0.975 8.The results showed that the performance of neural network method is better than that of MLR method.