脂肪族饱和一元醇的结构描述及沸点定量计算

On Structural Descriptors for Fatty Alcohols and Quantitative Structure Property Relationship (QSPR)

在烷烃分子距离边数矢量（ＭＤＥ） 基础上对分子中的（ － ＯＨ） 用染色因子进行了标识，提出了一种适用于描述脂肪醇分子结构的距离－ 边数（ＭＤＥ） 矢量μ．据此以１２４ 种脂肪醇分子的沸点为例，借助多元线性回归技术建立脂肪醇ＭＤＥ矢量与其沸点的定量结构－ 性质相关关系（ＱＳＰＲ） 模型，模型复相关系数达Ｒ＝ ０．９９４ ６，均方根误差为ξＲＭＳ＝ ３．２６ Ｋ．最后随机选取２５ 个化合物作预测集，以余下９９个作训练集再次建立ＱＳＰＲ模型（ Ｒ＝ ０ ．９９４ １，ξＲＭＳ＝ ３．４５ Ｋ） 并用它有效地预测了选定的２５ 个化合物的沸点（ Ｒ＝０ ．９９３ ７，ξＲＭＳ＝ ３．９９ Ｋ） ，结果表明，沸点的预测值与实验值的一致性很令人满意。

A new molecular distance edge (MDE) vector which can be applied to describe the structure of fatty alcohol molecules is proposed by using dying factors to describe the hydroxyl group ( OH) based on the molecular distance edge (MDE) vector of alkanes. The multiple linear regression (MLR) method is used to develop a model between the molecular distance edge (MDE) vector and boiling point (T B) of 124 fatty alcohol samples. The correlation coefficient and rout mean squares of the quantitative structure property relationship (QSPR) model are R =0.9946 and ξ RMS =3.26 K respectively. Then 5 predictive models are developed by using 99 fatty alcohols selected at random from the total 124 fatty alcohols as the training sets and their performances are demonstrated by employing 25 samples selected as the testing sets. And good results are obtained with the mean values of R and ξ RMS between the calculated T B value and experimental T B value being R=0.9941 and ξ RMS =3.45.All these results show that the predicted T B values are in good accordance with the experimental data.