The quantitative structure property relationship (QSPR) for thermal conductivity of liquid aliphatic alcohols was developed on the basis of 139 thermal conductivity data points of liquid aliphatic alcohols, which were divided into a 65-member training set, a 20-member validation set, and a 54-member prediction set. Four parameters (temperature-T, the intrinsic state pseudoconnectivity index-type 1s-Psi_i_1s, the sixth eigenvalue from augmented edge adjacency matrix weighed by edge degree-Eig06_AEA(ed), and the global topological charge index-JGT) were screened to develop the model by using the stepwise regression and the best subset regression method. For the training set, validation set and prediction set, the square correlation coefficient (R²) is 0.9769, 0.9726, and 0.9738, respectively. The mean relative deviation values of training set, validation set, and prediction set were 1.4%, 1.6%, and 1.6%. The QSPR model can provide not only basic data for the engineering application but also theoretical guidance for designing and seeking specific thermal conductivity materials.

中文翻译：

---

预测液态脂肪醇导热系数的改进的定量结构性质关系模型

基于液态脂族醇的139个热导率数据点，建立了液态脂族醇的热导率的定量结构性质关系（QSPR），将其分为65个成员训练集，20个成员验证集和54个成员的预测集。四个参数（温度-T，本征态伪连通性索引类型1s- Psi_i_1s，增强边缘邻接矩阵的第六个特征值以边缘度Eig06_AEA（ed）加权以及全局拓扑电荷指数-JGT使用逐步回归和最佳子集回归方法筛选）以开发模型。对于训练集，验证集和预测集，平方相关系数（R ²）分别为0.9769、0.9726和0.9738。训练集，验证集和预测集的平均相对偏差值为1.4％，1.6％和1.6％。QSPR模型不仅可以为工程应用提供基础数据，还可以为设计和寻找特定的导热材料提供理论指导。