基于原子类型电拓扑状态指数的 QSPR 模型预测链烷烃辛烷值

QSPR modeling with electrotopological state index for atom type for predicting octane number of paraffins

为了研究化合物结构与其辛烷值之间的相关性,以原子类型电拓扑状态指数作为描述符表征分子的结构特征,从40个样本中选取30种链烷烃作为训练集,对研究法辛烷值选取8个作为测试集,对马达法辛烷值选取7个作为预测集,利用SPSS软件进行多元线性回归,建立了辛烷值的定量结构-性质关系。得到的研究法辛烷值和马达法辛烷值预测模型复相关系数分别为0.965、0.955和均方根误差分别为5.4338、4.9404,并对模型进行内部和外部验证。结果表明,该模型具有较好的稳定性及预测能力。

In order to study the structure-octane number relationship of compounds, atom-type electrotopological state indices were used as molecular structure descriptors. The dataset of 40 paraffins was randomly divided into a training set containing 30 paraffins, a testing set consisting of 8 paraffins for research octane number and a testing set with 7 paraffins for motor octane number. SPSS software was employed to process multiple linear regression. A prediction model based on quantitative structure-property relationship （QSPR） studies was proposed, and the correlation coefficient and root-mean-square error of the model were 0.965 and 5.4338 for RON, 0.955 and 4.9404 for MON, respectively. Both internal and external validations were performed to validate the performances of the resulting model. The results showed that the model has good stability and prediction ability.