有机磷农药分子空间坐标指数与色谱保留值的QSPR研究

根据分子空间坐标数据,定义了原子距离指数Ys、分子空间距离指数yF和空间坐标特征值指数Yλ来表征有机磷农药（OPS）的分子结构特征;运用误差反向传输（BP）人工神经网络方法建立21种OPS空间坐标指数与气相色谱保留值定量结构性质分析（QSPR）模型,预测了气相色谱保留值.结果得出实验室值与预测值的相关系数R=0.950 2,所建模型稳定性和预测能力均良好.

Study of QSPR for Molecular Coordinates and Gas Chromatographic Relative Retention Time of Organophosphorus Pesticide

According to the molecular coordinates data, in order to characterization of the molecular structure characteristics of organ phosphorus pesticide that defines the atomic distance Ys, molecular space distance index YF and space coordinates Eigen value index Ya. This paper to forecast the gas chromatographic retention values, using back propagation （BP） artificial neural network method to establish 21 organphosphorus pesticides space coordinates index and gas chromatography retention Quantitative Structure Property Relationship （QSPR） model, the correlation coefficient of experimental value and predictive value is 0.950 2 and the model stability and prediction ability are good.