A new quantitative structure-property relationship (QSPR) model was developed to predict critical micelles concentration (CMC) of sugar-based surfactants using the norm indexes descriptors our group previously proposed. Statistics results showed that this model could provide satisfactory prediction results with the squared correlation coefficient for the whole dataset, the training set and the test set of 0.9545, 0.9547 and 0.9699 respectively. Several validation procedures including cross validation, Y-randomized test, mean absolute error test and applicability domain analysis as well as the comparison results with other works further demonstrated the stability and robustness of this model. Thus, the QSPR model using the norm indexes descriptors was effective in the prediction of CMC for sugar-based surfactants.

使用我们小组先前提出的标准指标描述符，开发了一种新的定量结构-性质关系（QSPR）模型来预测糖基表面活性剂的临界胶束浓度（CMC）。统计结果表明，该模型在整个数据集，训练集和测试集分别为0.9545、0.9547和0.9699的相关系数平方中可以提供令人满意的预测结果。包括交叉验证，Y随机测试，均值绝对误差测试和适用性域分析在内的几种验证程序以及与其他工作的比较结果进一步证明了该模型的稳定性和鲁棒性。因此，使用标准指标描述符的QSPR模型对于糖基表面活性剂的CMC预测是有效的。