酚类化合物对硝化颗粒污泥活性抑制的预测模型

摘要

Based on the molecular topological theory ,the molecular shape indices (mK) and atom-type electrotopo-logical state indices (En ) of 16 phenolic compounds are calculated .A four-variable model of the quantitative struc-ture-activity relationship (QSAR) model between the mK,En and the inhibition of phenolic compounds on activity of nitrifying granules is developed by using multiple linear regression and Leaps-and-Bounds regression .The traditional correlation coefficient R2 and the cross-validation correlation coefficient R2cv of leave-one-out is 0.932 and 0 .874 re-spectively .The QSAR model has both favorable estimation stability and good prediction capability by R2 ,F ,R2Adj , R2cv ,VIF,FIT tests .%基于拓扑理论计算了16种酚类化合物的分子形状指数（m K ）以及原子类型电性拓扑状态指数（ En ）．应用多元线性回归及最佳变量子集方法，建立了这些拓扑指数与酚类化合物对硝化颗粒污泥活性抑制的定量结构活性相关模型．该模型的传统判定系数 R2为0.932，逐一剔除法的交叉验证系数 R2cv为0.874，通过 R2，F ，R2Adj ，R2cv ， V IF ，FIT等检验，证明该模型具有良好的稳健性和预测能力．