论文部分内容阅读
定量结构-生物降解性能关系(QSBR)通过分析有机物结构与其生物降解性之间的定量关系,实现生物降解性的定量预测。针对影响生物降解性的基团结构多、传统方法难以消除基团数据之间的冗余,导致预测精度较低的问题,提出了一种基于主成分分析(PCA)-支持向量机(SVM)相结合的预测方法。首先利用主成分分析消除对该类化合物生物降解性影响较大的基团结构之间的冗余,降低数据维数,获取样本集主要信息;然后利用网格-交叉验证法优化后的支持向量机,建立预测模型。并与全要素的SVM模型及BP网络模型进行了比较,结果表明,该模型预测精度较高,具有通用性。
Quantitative Structure-Biodegradability Relationship (QSBR) Quantitative prediction of biodegradability is achieved by analyzing the quantitative relationship between the structure of an organic compound and its biodegradability. In order to solve the problem of low prediction accuracy, it is difficult to eliminate the redundancy between groups of data due to the large number of groups that affect biodegradability. A PCA-SVM (Support Vector Machine) A combination of prediction methods. First, the principal component analysis is used to eliminate the redundancy among the group structures that greatly affect the biodegradability of the compounds, and the data dimension is reduced to obtain the main information of the sample set. Then, the optimized support vector Machine, build predictive model. Compared with the total factor SVM model and the BP network model, the results show that the model has high prediction accuracy and versatility.