论文部分内容阅读
采用取代基片段值P和原子类型电拓扑状态指数Em有效表征了135个多氯咔唑化合物(PCCZs)的分子结构,通过选择变量与神经网络(BP)算法建立定量相关(QSPR)模型,以预测多氯咔唑化合物热力学性质.将选择的P,Em结构参数作为神经网络的输入层变量,热力学性质作为输出层变量,方程均采用5∶13∶1的网络结构,利用BP算法获得了3个令人满意的QSPR模型,它们的总相关系数分别为0.998 6,0.991 1和0.979 5,标准误差分别为2.123,3.237和3.952,利用这3个神经网络模型计算得到的预测值与文献值的相对平均误差分别为0.30%,1.85%和1.14%,表明模型具有良好的稳定性和预测能力.该神经网络模型所得结果优于多元回归方法所得结果,可用于对多氯咔唑化合物性质进行理论分析和预测.
The molecular structure of 135 polychlorocarbazole compounds (PCCZs) was characterized by the substituent fragment value P and the atomic topology index Em. The quantitative correlation (QSPR) model was established by choosing variables and neural network (BP) Predict the thermodynamic properties of polychlorocarbazole compounds.Using the P and Em structure parameters as input layer variables and thermodynamic properties as output layer variables of the neural network, the 5:13:1 network structure is used in the equations and the BP algorithm is used to obtain 3 A satisfactory QSPR model with a total correlation coefficient of 0.998 6, 0.991 1 and 0.979 5, respectively, with standard errors of 2.123, 3.237 and 3.952, respectively. Using these three neural network models, the predicted values and the literature values The relative mean errors are 0.30%, 1.85% and 1.14%, respectively, which shows that the model has good stability and predictive ability.The results obtained by this neural network model are superior to the results obtained by the multiple regression method and can be used to carry out theoretical analysis on the properties of polychlorocarbazole compounds Analysis and forecasting.