新分子电性距离矢量用于苯取代衍生物QSRR的研究  

Hydrogen-association classified molecular electronegativity-distance vector for quantitative structure-retention relationship of benzene derivatives

作　　者：杨国臣[1] 张生万[1] 仝建波[1] 

机构地区：[1]山西大学化学化工学院

出　　处：《计算机与应用化学》2007年第3期380-386,共7页Computers and Applied Chemistry

基　　金：山西省工业攻关项目基金[2006031204];山西省首届青年拨尖创新人才专项基金资助

摘　　要：本实验室新近提出按氢分类分子电性距离矢量(H-MEDV),用于对110个烷基苯气相色谱保留指数和54个二取代苯液相色谱容量因子进行结构表征,运用多元线性回归(multiple linear regression,MLR)以建立定量结构色谱保留关系(QSRR)模型,同时利用逐步回归结合统计检测对模型变量作筛选,建模计算值复相关系数(R_(cum))、留一法(leave-one-out,LOO)交互校验(cross-validation,CV)复相关系数(Q_(LOO)),对上述样本分别为0．9950、0．9937和0．9648、0．9530。研究结果表明,H-MEDV能较好表征该类分子结构信息,值得进一步推广。A newly developed hydrogen-association classified molecular electronegativity-distance vector （H-MEDV） was used to describe the chemical structure of benzene derivatives. Various multiple linear regression （MLR） models were created with variable screening by the stepwise multiple regression （SMR） technique. The correlation coefficients （Rcum） of the built models and Leave- One-Out （LOO） Cross-Validation（CV）, were Rcum = 0. 9950, QLOO= 0. 9937 for 110 alkyl benzenes and Rcum = 0. 9648, QLOO = 0. 9530 for 54 disubstituted benzenes, respectively. The models had favorable estimation stability and good prediction capabilities. Satisfactory results showed that information related to retention data of benzene derivatives could preferably be expressed by H-MEDV. H- MEDV can be a useful structural expression technique for quantitative structure retention （ activity or property） relationships （ QSRR/ QSAR/QSPR） and which may be applicable to the molecular structural characterization and chromatogram retention prediction.

关 键 词：按氢分类分子电性距离矢量 苯衍生物 色谱保留参数预测 定量结构-保留相关(QSRR) 

分 类 号：O641[理学—物理化学]