论文部分内容阅读
采用原子电性距离矢量(Atomic electronegative distance vector,AEDV),描述了14个有机酸类化合物中不同等价碳原子的化学微环境,以多元线性回归(MLR)和偏最小二乘回归(PLS)方法建立~(13)C核磁共振定量结构波谱关系模型,两模型的复相关系数(R)均为0.984和0.962,标准误差(SD)分别为11.021和16.244。经“留一法”交互检验的复相关系数(R_(CV))分别为0.981和0.939,标准误差(SD_(CV))分别为12.061和17.121。研究结果表明,使用该方法所建模型可以用于预测有机酸类化合物~(13)C NMR化学位移,模型具有良好的稳定性和预测能力。
The chemical microenvironments of different equivalent carbon atoms in 14 organic acids were described by Atomic electronegative distance vector (AEDV). Multiple linear regression (MLR) and partial least squares regression (PLS) Methods The ~ (13) C NMR quantitative structure spectral model was established. The complex correlation coefficients (R) of the two models were 0.984 and 0.962, respectively, and the standard deviations (SD) were 11.021 and 16.244, respectively. The correlation coefficient (R_ (CV)) of the “test-one-stay” method for cross validation were 0.981 and 0.939, respectively, and the standard errors (SD_ (CV)) were 12.061 and 17.121 respectively. The results show that the proposed model can be used to predict the ~ (13) C NMR chemical shifts of organic acids. The model has good stability and predictability.