论文部分内容阅读
通过对46个脂肪醇在不同固定相不同柱温下的631个样本的气相色谱保留指数值(RI)与其部分参数:拓扑指数(mQ)、定位基参数(Sox)、固定液极性值(CP)及柱温(T)建立定量结构-色谱保留相关(QSRR)模型。分别利用多元线性回归(MLR)、偏最小二乘回归(PLSR)、人工神经网络(ANN)建模,同时采用内部及外部双重验证的办法对所得模型稳定性能进行深入分析和检验,建模计算值、留一法(LOO)交互检验(CV)预测值和外部样本预测值的复相关系数Rmax、QLOO和Rexi分别为0．988,0．987和0．989(MLR);0．988、0．975和0．989(PLSR);0．989、0．989和0．991(ANN)。结果表明:所建定量结构保留相关(QSRR)模型具有良好的稳定性和预测能力,较好地揭示了脂肪醇化合物在不同固定相不同柱温上气相色谱保留指数的变化规律。
The gas chromatographic retention index (RI) of 631 samples of 46 fatty alcohols with different column temperatures at different stationary phases were compared with some of their parameters: topological index (mQ), positioning base parameter (Sox), fixed liquid polarity CP) and column temperature (T) to establish a quantitative structure-chromatographic retention correlation (QSRR) model. MLR, PLSR and ANN are respectively used to analyze and verify the stability of the model. The modeling and calculation are carried out by using both internal and external double verification methods. The correlation coefficients (Rmax, QLOO and Rexi) between the predictive value of LOO and the predicted value of external samples were 0.988,0.987 and 0.989 (MLR), respectively; 0.988, 0.975 and 0.989 (PLSR); 0.989, 0.989 and 0.991 (ANN). The results showed that the established QSRR model has good stability and predictive ability, and better reveals the change of GC retention index of fatty alcohol compounds at different column temperatures.