摘要
应用G98W程序包中的AMl方法对45个卤代甲烷分子进行优化计算,所得中心碳原子的静电荷密度(Qc)、分子的最高占据轨道能级(HOMO)和最低空轨道能级(LUMO)量化参数作为径向基人工神经网络的输入向量,选用df=10,eg=10-3对卤代甲烷ΔfHθm进行建模,所建模型预测值与其实验值吻合得很好.故这一新方法具有一定的参考价值.
The quantificational parameters which obtained from the optimize calculation to 45methane molecular by AM1 method in G98w software package are used as the inputs of the RBFnetwork. The model was established by choosing display - frequency = 10 and error - goa = 10^-3. There has excellent conformability between the value calculated by the model and the experimental data. Therefore, there＇s some referenced value on a certain degree.
出处
《青海师范大学学报（自然科学版）》
2005年第4期67-70,共4页
Journal of Qinghai Normal University(Natural Science Edition)
