摘要
为了研究脂肪醇的构效关系,定义并计算了脂肪醇中原子的特征值fi,利用量子化学AM1算法和分子图形学技术获得了分子拓扑参数和氢谱指数H,基于多元回归技术获得了对脂肪醇味阀值作出精确估算的定量结构-性质相关关系,得到多元回归方程,对50个脂肪醇分子味阀值估算的平均误差为0.46。利用方程对另外49个脂肪醇分子的味阀值进行预测,预测的平均误差为0.67,估算值、预测值与实验值较为吻合。
Atomic characteristic value fi of aliphatic alcohols were defined and calculated for QSPR research of aliphatic alcohols,then topological parameter and hydrogen line index H were acquired by the technique of molecule graphics and AM1. QSPR for estimating olfactory thresholds (POL) of aliphatic alcohols were achieved based on multiple regression. The regression equations were proposed. The mean error between estimated values and experimental value of 50 aliphatic alcohols is 0.46, while the olfactory thresholds of other 49 aliphatic alcohols molecules were predicted by the regression equation, the mean error is 0.67. The results show that the model developed can make a better agreement among estimated values, predicted values and experimental value.
出处
《石油化工高等学校学报》
CAS
2009年第3期28-32,共5页
Journal of Petrochemical Universities
基金
江苏省"青蓝工程"科研基金资助项目(QL20072)
徐州工程学院重点科研基金资助项目(XKY2007105)