论文部分内容阅读
本文利用主成分提取-线性判别分析(PCA-LDA)模型对多环芳烃(Polycyclic Aromatic Hydrocarbons,PAHs)的致癌性进行分类,与致癌性有关的多环芳烃的表面积、代谢活性区域中心碳原子的离域能、亲电活性区域中心碳原子的离域能以及分子脱毒区的总数四个参数作为模型的输入,用已知致癌性的67个样本作为训练集建立PCA-LDA模型,对10个预测集样本的致癌性进行预测,结果表明:致癌性按高(h)、低(l)、非(n)分类时预测准确率达100%。
In this paper, the principal component extraction-linear discriminant analysis (PCA-LDA) model was used to classify the carcinogenicity of polycyclic aromatic hydrocarbons (PAHs), the surface area of PAHs, the carbon atoms The four parameters of delocalization energy, the delocalization energy of the center of the electrophilic active region and the total number of molecular detoxification regions were used as the input of the model. Sixty-six samples of known carcinogenicity were used as a training set to establish the PCA-LDA model. The predicted carcinogenicity of 10 predicted samples showed that the predicted accuracy of carcinogenicity was 100% when classified by high (h), low (l), and non (n).