论文部分内容阅读
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。
A quantitative structure-activity relationship (QSAR) model was established between the structure of 16 fluorinated phenols and their toxicity to Tetrahymena spp. Using error feedback forward artificial neural network (ANN). Sixteen kinds of fluorinated phenols were used as input for quantum chemistry and physicochemical parameters. The acute toxicities of Tetrahymena spp. Were taken as output. The stability and external predictive ability of the obtained model were verified by both internal and external validation methods. The network model The correlation coefficient was 0.999 8, the correlation coefficient of cross test was 0.981 8, the standard deviation was 0.01 and the residual absolute value was less than 0.04, which was applied to the external forecasting set. The correlation coefficient of the external forecasting set was 0.993 6. The multiple linear regression (MLR) The model has a correlation coefficient of 0.980 2, a standard deviation of 0.119, an absolute residual value of ≤0.28 and a correlation coefficient of 0.980 3 for the external prediction set. The results show that ANN model has better fitting effect than MLR model.