采用以基团间距离、边数为基准的分子距边矢量作为二(口恶)(口英)类化合物多氯代二苯并呋喃(PCDFs)的分子结构描述符,结合反向传播人工神经网络方法建立了PCDFs的正辛醇/水分配系数(logKow)与分子结构描述符之间的定量关系模型。结果表明,所采用的结构描述符对于分子结构具有很好的区分能力,所建立的模型对于校验样本的预测精度较高。利用所得模型,对文献尚未报道logKow实验值的其他所有85种PCDFs给出了预测值。 Using the molecular distance vector based on the distance between groups and the number of edges as the descriptor of the molecular structure of PCDFs and combining with the back propagation artificial nerve The network method establishes a quantitative relationship model between the log-normal octanol / water partition coefficient (PCK) and molecular structure descriptors of PCDFs. The results show that the structure descriptors used in this paper have good discriminating ability for the molecular structure, and the established models have higher predictive accuracy for the calibration samples. Using the model obtained, predictions for all 85 other PCDFs that have not been reported in the literature for logKow values ​​are given.