构建了基于分子三维结构计算的原子电负性距离矢量(atomic electronegative space distance vector,AESDV),用以描述各芳香醚类化合物分子中不同等价碳原子的化学微环境,并结合原子自身杂化状态指数(AHSI),建立了^(13)C核磁共振定量结...构建了基于分子三维结构计算的原子电负性距离矢量(atomic electronegative space distance vector,AESDV),用以描述各芳香醚类化合物分子中不同等价碳原子的化学微环境,并结合原子自身杂化状态指数(AHSI),建立了^(13)C核磁共振定量结构波谱关系的多元线性回归模型,复相关系数(R)为0.964,标准误差(SD)为8.673.经留一法交互检验的复相关系数(R_(CV))为0.948,标准误差(SD_(CV))为10.362.随机抽出样本进行外部检验,得到测试集的复相关系数(R_(test1)及R_(test2))分别为0.979和0.939,标准误差(SD_(test1)及SD_(test2))分别为6.400和10.162.研究结果表明,使用该方法所建模型具有良好的预测能力和稳定性.展开更多
Molecular structures of reactants were characterized by molecular electronegativity distance vector （VHMED） considering hydrogen association. A reasonable molecular modeling equation with 4-parameters was achieved f...Molecular structures of reactants were characterized by molecular electronegativity distance vector （VHMED） considering hydrogen association. A reasonable molecular modeling equation with 4-parameters was achieved for quantitative structure-property/activity relationship （QSPR/QSAR） by stepwise multiple regression （SMR） that the variable was introduced item by item in significant level order. A high correlation coefficient （R = 0.980） demonstrates that the model is able to well express a quantitative relation between stereoselectivity and the reactant structures as quantitative structure-reactivity/stereoselectivity relationship （QSRR/QSSR）. The multiple correlation coefficient （Rcv= 0.964） was tested through cross-validation with the leave-one-out （LOO） procedure. The above results show that the model possesses high estimation stability and good prediction ability between the amount of both cis and trans isomers in products and reactants.展开更多
基金Youth Foundation of Education Bureau,Sichuan Province(13ZB0003)Natural Science Foundation of Education Bureau,Sichuan Province(15ZB0272)
文摘构建了基于分子三维结构计算的原子电负性距离矢量(atomic electronegative space distance vector,AESDV),用以描述各芳香醚类化合物分子中不同等价碳原子的化学微环境,并结合原子自身杂化状态指数(AHSI),建立了^(13)C核磁共振定量结构波谱关系的多元线性回归模型,复相关系数(R)为0.964,标准误差(SD)为8.673.经留一法交互检验的复相关系数(R_(CV))为0.948,标准误差(SD_(CV))为10.362.随机抽出样本进行外部检验,得到测试集的复相关系数(R_(test1)及R_(test2))分别为0.979和0.939,标准误差(SD_(test1)及SD_(test2))分别为6.400和10.162.研究结果表明,使用该方法所建模型具有良好的预测能力和稳定性.
文摘Molecular structures of reactants were characterized by molecular electronegativity distance vector （VHMED） considering hydrogen association. A reasonable molecular modeling equation with 4-parameters was achieved for quantitative structure-property/activity relationship （QSPR/QSAR） by stepwise multiple regression （SMR） that the variable was introduced item by item in significant level order. A high correlation coefficient （R = 0.980） demonstrates that the model is able to well express a quantitative relation between stereoselectivity and the reactant structures as quantitative structure-reactivity/stereoselectivity relationship （QSRR/QSSR）. The multiple correlation coefficient （Rcv= 0.964） was tested through cross-validation with the leave-one-out （LOO） procedure. The above results show that the model possesses high estimation stability and good prediction ability between the amount of both cis and trans isomers in products and reactants.
