【摘要】 为了研究人工神经网络预测有机化合物生物降解的性能，运用多元线性回归方法和误差反向传递人工神经网络模型以基团代码作为结构描述符，分别拟合、预测了一批含硫芳香族化合物的一级好氧生物降解速率常数．结果表明，由于神经网络自动考虑了基团间的交互作用，它对生物降解这类复杂问题有极高的求解能力，预测的均方误差为０．００１０２，远低于线性回归模型的预测误差０．０１５９１更多还原

【Abstract】 In order to evaluate the ability of artificial neural networks to predict the biodegradability of organic compounds,the codes of function groups were employed as structural descriptors to fit and predict the pseudofirstorder aerobic biodegradation rate constants of a batch of sulfurcontaining aromatic compounds using multiple linear regression and error back propagation neural network models,respectively.The results indicated that the neural network has higher ability to solve the complex problems affected by confounding structure factors such as biodegradation process due to its ability to conclude interactions between groups automatically.The meansquare error of the network is 000102,which is much lower than that of the linear model 001591.更多还原