人工神经网络用于氟化酚的定量构效关系研究  

QSAR Study of Fluorinated phenols by ANN

作　　者：何琴[1] 闫玲玉[1] 黄保军[2] 

机构地区：[1]许昌学院化学化工学院,河南许昌461000 [2]河南省微纳米能量储存与转换材料重点实验室,许昌学院表面微纳米材料研究所,河南许昌461000

出　　处：《河北化工》2012年第6期28-31,共4页Hebei Chemical Industry

基　　金：河南省教育厅自然科学研究计划项目(批准号:2009B150023)

摘　　要：采用误差反传前向人工神经网络(ANN)建立了16种氟化酚的结构与其对梨形四膜虫的毒性之间的定量结构-活性关系(QSAR)模型。以16种氟化酚的量子化学和理化参数作为输入,对梨形四膜虫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和外部预测能力,所构建网络模型的相关系数为0.999 8、交叉检验相关系数为0.981 8、标准偏差为0.01、残差绝对值≤0.04,应用于外部预测集,外部预测集相关系数为0.993 6;而多元线性回归(MLR)法模型的相关系数为0.980 2、标准偏差为0.119、残差绝对值≤0.28,外部预测集相关系数为0.980 3。结果表明,ANN模型获得了比MLR模型更好的拟合效果。To set up the quantitative structure-activity relationship（QSAR） model on 16 fluorinated phenols by the artificial neural network（ANN） based on the back propagation algorithm.For the ANN method,when using the quantum chemical and physical chemical parameters about structure as the inputs of the neural network and the acute toxicities as the outputs of the neural network,the correlation coefficient was 0.999 8,the leave one out cross-validation regression coefficient was 0.981 8,the standard error was 0.01,the correlation coefficient of the test set was 0.993 6 and the absolute values of residual were less than 0.04.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 0.980 2,the standard error was 0.119,the absolute values of residual were less than 0.28 and the correlation coefficient of the test set was 0.980 3.The results showed that the performance of neural network method was better than that of MLR method.

关 键 词：氟化酚 定量构效关系 人工神经网络 急性毒性 

分 类 号：O625.1[理学—有机化学]