论文部分内容阅读
有机化合物通常由电性各异的氢、碳、氮、氧、硫及卤素等原子以共价键结合而成,从分子的三维空间结构出发,按照不同类型原子之间电性相互作用得到一种分子结构表达方法——三维分子电性相互作用矢量(three dimensional molecular electronegativity interaction vector,3D-MEIV).利用该矢量对41种烃基酚类化合物进行结构表达,通过逐步回归的方法建立了2变量毒性模型,复相关系数(R)为0.9438,均方根误差(RMS)为0.2541.再用留一法(Leave-one-out,LOO)交互检验对模型进行了评价,得到的复相关系数(R(CV))为0.9349,均方根误差(RMS(CV))为0.2728.结果表明3D-MEIV能较好地表征烃基酚类化合物的分子结构,所建模型具有较好的稳定性和预测能力.
Organic compounds are usually covalently bonded by atoms of different types such as hydrogen, carbon, nitrogen, oxygen, sulfur and halogen. From the three-dimensional structure of the molecule, the organic compounds are obtained according to the electrical interaction between different types of atoms Three dimensional molecular electronegativity interaction vector (3D-MEIV) was used to express the structure of 41 kinds of hydrocarbyl phenols, and two variables were established by stepwise regression Toxicity model, the complex correlation coefficient (R) was 0.9438, and the root mean square error (RMS) was 0.2541. The model was evaluated by leave-one-out (LOO) R (CV)) was 0.9349, and the root mean square error (RMS (CV)) was 0.2728. The results showed that 3D-MEIV could better characterize the molecular structure of alkylphenols and the model has good stability and prediction ability.