论文部分内容阅读
基于定量结构-色谱保留相关(QSRR)研究醇类化合物的保留性质具有重要意义。通过对脂肪醇的研究,考虑到烷基和羟基的差异,修正文献,重新定义脂肪醇分子中顶点原子特征值t_i,并在邻接矩阵的基础上构建新的连接性指数~mL。运用多元线性回归法,将0阶指数~0L、1阶指数~1L与25种脂肪醇,在6种固定相(SE-30,OV-3,OV-7,OV-11,OV-17和OV-25)上的气相色谱保留指数RI回归分析,发现~0L、~1L与RI显著相关,建模的复相关系数R分别达到0.999 3、0.999 6、0.998 8、0.9977、0.997 2和0.997 0,均优于前人的工作。继用留一法和外部样本检验模型(SE-30)的稳定性和预测能力,结果表明:交互检验复相关系数R_(CV)在0.999 3～0.999 4之间,模型稳定性很好;在固定相SE-30上,10种醇的预测值与实验值的平均相对误差仅为0.87%,二者的一致性令人满意。
It is of great importance to study the retention properties of alcohols based on QSRR. Through the study of aliphatic alcohols, taking into account the differences between alkyl and hydroxyl groups, the literature is revised to redefine the eigenvalue t_i of the vertex atoms in the fatty alcohol molecule and to construct a new connectivity index ~ mL based on the adjacency matrix. Using the multiple linear regression method, the first order index (0L), the first order index (1L) and 25 kinds of fatty alcohols were detected in 6 stationary phases (SE-30, OV-3, OV-7, OV-11, OV-25). The regression coefficients of ~ 0L and ~ 1L were significantly correlated with RI. The complex correlation coefficients R of the model were 0.999 3, 0.999 6, 0.998 8, 0.9977, 0.997 2 and 0.997 0 respectively , Are better than their predecessors work. The stability and prediction ability of the SE-30 model were tested by using the left-behind method and the external sample test. The results showed that the complex correlation coefficient R_ (CV) was between 0.999 3 and 0.999 4. The stability of the model was good. On the stationary phase SE-30, the average relative error between the predicted and experimental values of the 10 alcohols was only 0.87%, and the agreement between them was satisfactory.