摘要
用神经网络反向传播算法计算了双原子分子的键长。采用二原子的 Slater 原子半径,Paul-ing 电负性,在元素周期表中的主族数及周期数等作为特征变量,得到了神经网络的训练及预报结果。
By using the back-propagation(BP)model of neural network,the bond length of diatomic molecule is determinated.Slater atomic radius,Pauling electronegativity,major group and cycle numbers of the periodic table,etc.,are used as features.The calculated and predicted results of neural network appear to be better than those of reported calculations.
出处
《计算物理》
CSCD
北大核心
1996年第2期243-248,共6页
Chinese Journal of Computational Physics
关键词
神经网络
双原子
分子
键长
neural network
diatomic molecule
bond length
electronegativity
features
