论文部分内容阅读
本文从文献中收集了多个钙钛矿结构的掺杂LaGaO_3系列氧离子导体电解质材料样本,以导电率的对数Ln(?)为目标,使用各种机器学习方法进行回归分析,包括多元线性回归(MLR)、偏最小二乘法(PLS)和支持向量回归(SVR),建立了Ln(?)与其分子结构参数之间的定量模型。结果表明:SVR方法所得导电率Ln(?)的留一法预报结果与实验最相符,计算值与实验值的相关系数为0.911。使用独立测试集预报的计算值和实验值的相关系数为0.880。此外还用建立的模型对La_(1-x)Sr_xGa_(1-y)Mg_yO_3掺杂体系的导电率进行了预报,根据预报结果做出的等高面图显示的优区与实验所得结果一致。
In this paper, we collected a series of perovskite-doped LaGaO 3 series oxygen ion conductor electrolyte materials samples from the literature, with the logarithm Ln (?) Of conductivity as a target, using various machine learning methods for regression analysis, including multiple linear (MLR), partial least squares (PLS) and support vector regression (SVR), a quantitative model between Ln (?) And its molecular structure parameters was established. The results show that the predictive value of the Ln (?) Conductivity obtained by the SVR method is in good agreement with the experimental one, and the correlation coefficient between the calculated and experimental values is 0.911. The correlation coefficient between the calculated and experimental values predicted using the independent test set was 0.880. In addition, the conductivity of La 1-x Sr x Ga 1- (1-y) Mg_yO 3 -doped system was also predicted by the established model. The results of the contour plots made by the prediction results are consistent with the experimental results.