基于启发式方法和支持向量机方法预测有机物的热导率

构建147个有机物分子结构与其热导率值之间的定量结构-性质关系(QSPR)模型, 探讨影响有机物热导率的结构因素. 以147个化合物作为样本集, 随机选择118个作为训练集, 29个作为测试集. 应用CODESSA软件计算了组成、拓扑、几何、静电和量子化学等描述符, 通过启发式方法(HM)筛选得到5个结构参数并建立线性回归模型; 用所选5个结构参数作为支持向量机(SVM)的输入, 建立非线性的支持向量机回归模型. 预测结果表明: 支持向量机回归模型的性能(复相关系数R²=0.9240)虽略低于启发式回归模型的性能(R²=0.9267), 但是支持向量机方法预测性能(R²=0.9682)高于启发式方法的预测性能(R²=0.9574), 对于QSPR模型来说, 预测性能更重要. 因此, 总体来说支持向量机方法优于启发式方法. 支持向量机方法和启发式方法的提出为工程上提供了一种根据分子结构预测有机物热导率的新方法.     

Estimation and Prediction of Bioconcentration Factors of Nonionic Organic Chemicals in Fish by Electrotopological State Indices and Structural Parameter

Based on the characteristics of atom types, Hall＇s electrotopological state indices （En） are calculated for 165 nonionic organic compounds. On the basis of the characteristics of substituent and conjugated matrix, a novel molecular structure parameter （G） is defined and calculated for 165 molecules in this paper. En and G show good structural selectivity for organic molecules. G, a satisfactory relationship between bioconcentration factor （BCF） and En, is expressed as： 1gBCF = -0.283 ＋ 1.246G ＋ 0.079E42 ＋ 0.351E9- 0.063E17 （n＇ = 122, R = 0.967, F = 425.636, s = 0.394）, which could provide estimation and prediction for the lgBCF of nonionic organic chemicals. Furthermore, the model is examined to validate overall robustness with Jackknife tests, and the independent variables in model do not exist cross correlation with VIF. All these regression results show that the new parameter G and electrotopological state index have good rationality and efficiency. It is concluded that the En and G will be used widely in quantitative structure-property/activity relationship （QSPR/QSAR） research.