论文部分内容阅读
采用分子子图编码法将烷烃的分子子图码作为人工神经网络的输入参数 ,对烷烃的吉布斯自由能进行预测 .将预测值与文献值进行线性拟和 ,其回归系数达到 0 9812 .
Molecular code of alkane was used as the input parameter of artificial neural network to predict the Gibbs free energy of alkane by molecular sub-code method. The predicted value and literature value were linearly fitted and the regression coefficient reached 0 9812.