论文部分内容阅读
基于分子全息技术, 研究了 135 种多氯代二苯并呋喃化合物(PCDFs)气相色谱保留指数(GC-RI)与分子全息结构之间的关系, 应用偏最小二乘(PLS)回归技术建立了分子全息定量结构-色谱保留(QSRR)模型. 交叉验证相关系数qLOO = 0.998, 非交叉验证系数 r2 = 0.998. 随机选出 100 种 PCDFs 化 2合物作为训练集, 其余作为测试集, 来验证分子全息 QSRR 模型的预测能力和稳健性. 采用最佳模型的建模条件对测试集进行 PLS 分析, qLOO 为 0.997, r2为 0.998. 用训练集数据所建立的 QSRR模型预测测 2试集中 PCDFs 化合物的色谱保留指数, 结果表明基于训练集所建立的 QSRR 模型可以对测试集中PCDFs 化合物的气相色谱保留指数进行很好的预测.
Based on molecular holography, the relationship between the gas chromatographic retention index (GC-RI) and the molecular holographic structure of 135 polychlorinated dibenzofuran compounds (PCDFs) was studied. Partial least square regression (PLS) The molecular holographic quantitative structure-chromatographic retention (QSRR) model with cross-validated correlation coefficient qLOO = 0.998 and non-cross-validation coefficient r2 = 0.998 randomly selected 100 PCDFs as the training set and the rest as the test set to validate the molecular The prediction ability and robustness of the holographic QSRR model PLS analysis of the test set was performed using the best model modeling conditions with qLOO of 0.997 and r2 of 0.998.The QSRR model established by training set data was used to predict the concentration of PCDFs Chromatographic retention index, the results show that based on training set established QSRR model GCDFs compounds can be tested on the GC retention index for a good prediction.