烷基硫醇在不同固定相上的QSRR研究

Studies on quantitative structure-retention relationship for sulfur alcohol on different stationary phases

计算出体系中烷基硫醇化合物的各种结构参数,以优选出的分子连接性指数和量化参数为结构描述符,首次采用反向传播算法(BP)人工神经网络、径向基函数网络(RBF)2种非线性方法建立了参数少且精度高的定量结构-色谱保留相关(QSRR)模型,预测了烷基硫醇在4种极性固定上的气相色谱保留指数(RJ)。结果表明:在4种固定相上建立的BP模型均优于RBF模型且非线性方法(BP、RBF)优于文献中多元线性回归(MLR)方法,所建定量结构保留关系(QSRR)模型具有良好的稳定性和预测能力。

Solute-related parameters of the sulfur alcohol compounds were calculated by molecule mechanics and quantum chemistry method. A stepwise multiple linear regression procedure carried out by SPSS was used to select variables which were related to the target values. The intermolecular interaction index and quantum chemical parameters were selected as the inputs of the Back Propagation （BP） and Radial Basis Function （RBF） artificial neural network models. The results performed well indicating that these QSRR models have an excellent stability and predictive capability for predicting retention indices of sulfur alcohol compounds. Also, obtained results indicated that both BP and RBF methods have achieved better predictive performance than MLR method that taken from reference. Contrasting RBF model, the predictive performance of BP model still achieve a little improvement.