论文部分内容阅读
基于对著名的Kier连接性指数(mXtv)修正,定义了一种新型分子连接性指数—Kier逆指数(mKtv,)并计算了木香花挥发油中46种有机成分的Kier逆指数值。使用最佳变量子集回归(LBR)建立了46种有机成分的气相色谱保留时间(tR,min)与其Kier逆指数的三元数学模型,其传统相关系数(R)为0.930,逐一剔除法(LOO)交叉验证系数(Q)为0.914。该模型具有高度稳健性与良好的预测能力,影响有机物气相色谱保留时间的主要因素是分子大小与空间形状。依此3个Kier逆指数作为神经网络输入层结点,采用3∶5∶1的网络结构,利用BP算法得到一个良好模型,其相关系数(R)为0.994,显示了两者之间具有良好的非线性关系。结果表明,Kier逆指数对有机物气相色谱保留时间的表征是合理的、有效的,所建模型能较好解释其递变规律。
Based on the famous modification of the Kier connectivity index (mXtv), a new molecular connectivity index, the Kier inverse index (mKtv,), was defined and the Kier inverse index values of 46 organic components in the volatile oil of the flowers were calculated. Using the optimal variable subset regression (LBR), a ternary mathematical model for the retention time (tR,min) and the Kier inverse index of 46 organic components was established. The traditional correlation coefficient (R) was 0.930, and the one-by-one elimination method was used ( LOO) The cross-validation coefficient (Q) is 0.914. The model has a high degree of robustness and good predictive ability. The main factors affecting the gas chromatographic retention time of organics are the size of molecules and the shape of space. According to this 3 Kier inverse exponent as the neural network input layer node, using a 3:5:1 network structure, using BP algorithm to get a good model, the correlation coefficient (R) is 0.994, showing that there is good between the two The nonlinear relationship. The results show that the Kier inverse exponent is reasonable and effective for the gas chromatographic retention time of organic compounds. The model can better explain the law of its evolution.